(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Bulletin of the British Museum (Natural History) Geology Supplement"

J-£/r) 



\OX) 



A REVISION OF THE ELOPIFORM 
FISHES, FOSSIL AND RECENT 



P. L. FOREY 



BULLETIN OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 
GEOLOGY Supplement 10 

LONDON: 1973 



u <2 



A REVISION OF THE ELOPIFORM FISHES, 
FOSSIL AND RECENT 



BY 

PETER L. FOREY 

Department of Biology, Queen Elizabeth College, 
University of London, London 



Pp 1-222 ; 92 Text-figures 



BULLETIN OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 
GEOLOGY SUPPLEMENT 10 

LONDON: 1973 



THE BULLETIN OF THE BRITISH MUSEUM 

(natural history), instituted in 1949, is 
issued in five series corresponding to the Departments 
of the Museum, and an Historical series. 

Parts will appear at irregular intervals as they 
become ready. Volumes will contain about three or 
four hundred pages, and will not necessarily be 
completed within one calendar year. 

In 1965 a separate supplementary series of longer 
papers was instituted, numbered serially for each 
Department. 

This paper is Supplement 10 of the Geological 
(Palaeontological) series. The abbreviated titles of 
periodicals cited follow those of the World List of 
Scientific Periodicals. 



World List abbreviation : 
Bull. Br. Mus. nat. Hist. (Geol.) Suppl. 



© Trustees of the British Museum (Natural History), 1973 



TRUSTEES OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 

Issued 21 December, 1973 Price £9.45 



A REVISION OF THE ELOPIFORM FISHES, 
FOSSIL AND RECENT 

By PETER L. FOREY* 



CONTENTS 

I. Introduction 
II. Materials and methods 

III. Systematic descriptions 

Cohort TAENIOPAEDIA 
Superorder Elopomorpha 
Order Elopiformes 
Suborder Elopoidei 
Family Elopidae Bonaparte 

Genus Elops Linnaeus 

Genus Davichthys gen. nov. 

Genus Anaetlialion White . 
Family Megalopidae Jordan 

Genus Tarpon Jordan & Evermann 

Genus Megalops Lacepede . 

Genus Protarpon gen. nov. . 

Genus Promegalops Casier . 

Genus Elopoides Wenz 

Genus Sedenhorstia White & Moy-Thomas 

Genus Pachythrissops Woodward. 
Suborder Albuloidei 
Family Osmeroididae nov. . 

Genus Osmeroides Agassiz . 

Genus Dinelops Woodward. 
Family Pterothrissidae Gill 

Genus Pterothrissus Hilgendorf 

Genus Istieus Agassiz 

Genus Hajulia Woodward . 
Family Albulidae Bleeker . 

Genus Albula Scopoli. 

Genus Lebonichthys gen. nov. 

IV. Discussion of the Order Elopiformes . 

(a) Historical ..... 

(b) Salient features of the order Elopiformes and relationships 

with other ' lower teleosts ' 

(c) Suborder Elopoidei 

(d) Suborder Albuloidei 

(e) Possible albuloid derivatives 

(f) Comments on the Family Phyllodontidae 
V. Summary ...... 

VI. Acknowledgements .... 

VII. References ..... 

VIII. Abbreviations used in the text figures 



Page 
4 
5 
6 
6 
6 
6 
7 
7 
7 
28 

35 
44 
44 
6 4 
70 
80 
82 

85 
88 

94 

94 

94 

112 

124 
141 

152 
153 
153 
171 

185 

185 

187 
196 
202 
211 
211 
212 
214 
214 
221 



* Present address: Department of Zoology, University of Alberta, Edmonton 7, Alberta, Canada. 



4 ELOPIFORM FISHES 

SYNOPSIS 

The skeletons of the Recent elopiform fishes (Elops, Tarpon, Megalops, Pterotlirissus, Albula) 
are described, and all fossil genera thought to be correctly placed in the Elopiformes are reviewed 
and compared with the living forms. From the descriptions it is concluded that the classification 
of the Elopiformes given by Greenwood et al. (1966) reflects the phylogeny of this group. A 
new family, the Osmeroididae, based on the Cretaceous genus Osmeroides, is included in the 
suborder Albuloidei as the basal family of that suborder. The Osmeroididae provide a link 
between the Elopoidei and the more specialized Albuloidei. Primitive and advanced features 
of the Elopiformes are discussed together with individual discussions of the contained suborders. 
Three new fossil genera are established : Davichthys, Protarpon and Lebonichthys. The relation- 
ship of the Elopiformes with other basal teleost groups is considered and it is concluded that 
although no relationship exists with the Clupeomorpha and Osteoglossomorpha within the 
Teleostei, the possibility of a relationship with the Euteleostei cannot be ruled out. 

I. INTRODUCTION 

This work is a taxonomic revision of the order Elopiformes Greenwood et al. (1966). 
The Elopiformes are a group of primitive teleosts related to the highly specialized 
Anguilliformes and Notacanthiformes by virtue of the fact that members of all 
three orders show a leptocephalus larval stage in their life history. The elopiforms 
are represented in the Recent fauna by a handful of genera ; Elops, Megalops, 
Tarpon, Albula, Dixonina and Pterothrissus. There are more fossil genera, found 
from the Upper Jurassic onwards. Woodward (1901) lists eighteen genera while 
Romer (1966) includes some forty-six in the Elopiformes. Although many of these 
fossil forms do not appear to be members of the Elopiformes, it remains true that the 
order is largely constituted by extinct genera. 

Most fossil elopiforms are known from Upper Cretaceous deposits and were 
originally described in several faunal works. The more important of these works 
are those of Agassiz (1833- 1844) in which forms from several localities were dealt 
with ; Dixon (1850) described the English Chalk species ; Pictet (1850), Pictet & 
Humbert (1866), Davis (1887) and Hay (1903) described those from the Lebanon. 
Marck (1858, 1863) noted forms from the Westphalian Chalk and Cope (1872), 
Loomis (1900) and Stewart (1898) were concerned with those from the Kansas 
Chalk. This earlier work was synthesized by Woodward (1901) who went on to 
deal extensively (1907, 1908) with the English Chalk fauna. In more recent years 
d'Erasmo (1946) has published on the Upper Cretaceous fauna from Comen, Yugo- 
slavia, Arambourg (1954) has described elopiforms from the Cretaceous of Morocco 
and a revision of the Westphalian fauna has been undertaken by Siegfried (1954). 

Important work on the anatomy and taxonomy of certain fossil genera has been 
published by Nybelin (1967b) and Gaudant (1968) for Anaethalion, and by Goody 
(1969a) for Sedenhorstia. Recent elopiforms have received attention in faunal 
works by Barnard (1925), Fowler (1936), Okada (i960) and Hildebrand (1963). 
Finally, important contributions to the anatomical features of the Recent forms 
have been made by Ridewood (1904), Hollister (1936, 1939), Nybelin (1956, 1967a, 
1971) and Greenwood (1970a). Apart from these specific references the genera Elops 
and Megalops have often been referred to in papers concerned with phylogeny 
and comparative anatomy. 



FOSSIL AND RECENT 5 

The starting point of the work was taken from Woodward (1901), who recognized 
the two families Elopidae and Albulidae. Woodward (1901) used a key to divide 
the Elopidae into two assemblages, later termed the ' elopine ' and ' spaniodontine ' 
groups by Dunkle (1940). This basic division of the elopids was based on the 
condition of the parietals, medially united in the ' elopines ' and separated by the 
supraoccipital in the ' spaniodontines '. During the course of this revision it became 
apparent that many of the ' spaniodontine ' genera together with two from the 
' elopine ' group represent an assemblage distinct from the Elopiformes. These 
genera (Notelops, Rhacolepis, Pachyrhizodus, Thrissopater, Esocelops and possibly 
Spaniodon and Thrissopteroides) differ from elopiform fishes in many important 
details of the snout, the otic region of the neurocranium, the jaws, the infraorbital 
series, the vertebral column and caudal skeleton, which show that these genera are 
more correctly placed with the euteleostean fishes. They will be dealt with in a 
later publication. 

Of the genera included by Woodward in the family Albulidae only Albula and 
Istieus are retained here. Chanoides, Chanos, Prochanos and probably Ancylostylos 
are all interrelated and their affinities lie with the Gonorynchiformes. Ananogmius 
White & Moy-Thomas (now. subst. for Anogmius Cope, 1877) shows many cranial 
and caudal features indicating that it is a member of the osteoglossomorph family 
Plethodontidae (for features of the plethodonts see Bardack 1965 and Patterson 
1967c). 

It has been impossible to examine at first hand all genera which have at one time 
or another been referred to the Elopiformes. Where relevant, comments are in- 
cluded on these unexamined forms, but the many records of isolated otoliths and 
scales have been omitted because of their debatable value in a work of this nature. 

Greenwood et al. (1966), writing from a neontological point of view, proposed a 
classification for the Elopiformes which is substantiated here. The Elopiformes 
are separable into two rather different suborders, the Elopoidei Jordan (1923) and 
the Albuloidei Greenwood et al. (1966). Following a general discussion of the order, 
the suborders are given separate consideration. 



II. MATERIALS AND METHODS 

Nearly all material used in the preparation of this work is in the British Museum 
(Natural History) and those specimens are referred to by the prefix B.M.N.H. A few 
specimens were borrowed from the Royal Scottish Museum, Edinburgh, and are 
prefixed by R.S.M. The prefix M.H.N. P. denotes specimens in the Museum 
National d'Histoire Naturelle, Paris. 

My thanks are due to Professor T. Abe of Tokyo University for the donation of 
three specimens of Pterothrissus gissu, one specimen of Elops hawaiensis and one of 
Megalops cyprinoides. 

Techniques used in the preparation of fossil and Recent material were those used 
by Goody (1969b). The clearing of specimens prior to staining with ' alizarin S ' 
was done using the enzyme-clearing technique described by Taylor (1967). 



6 ELOPIFORM FISHES 

The standard length (S.L.) of the specimens is taken from the tip of the snout to 
the anterior margin of the first ural centrum, unless otherwise stated. The frequent 
displacement of the hypurals in fossil material prevented the use of the more usual 
limits of this co-ordinate. The length of the head is interpreted as the distance 
between the tip of the snout and the posterior margin of the operculum. Linear 
dimensions are expressed in millimetres. Vertebral counts given are exclusive of 
ural centra and centra incorporated as a functional part of the neurocranium. 
Lastly, the first caudal vertebra is taken to be that centrum which bears a complete 
haemal arch and spine. 



III. SYSTEMATIC DESCRIPTIONS 

Cohort TAENIOPAEDIA Greenwood et al. (1967) 

Superorder ELOPOMORPHA Greenwood et al. (1966) 

Order ELOPIFORMES Greenwood et al. (1966) 

Diagnosis. Elopomorph fishes of fusiform shape ; body rounded or compressed, 
never excessively deepened or elongated. Dorsal and anal fins emarginate, always 
distinct from the forked caudal fin. Pectoral fins held horizontally, low down on 
the body ; pelvic fins abdominal. Epaxial trunk musculature never extending 
onto the cranial roof. Dermethmoid large, overlying a cartilaginous ethmoid. 
Parietals meeting in the mid-line. Post-temporal fossa with a roof ; subtemporal 
fossa deep ; pars jugularis long ; jugular vein, hyomandibular branch of VII and 
orbital artery with separate openings on the lateral face of the prootic. Orbito- 
sphenoid and basisphenoid well developed. Anterior myodome absent. Para- 
sphenoid extending to, or near the posterior limit of the neurocranium, with teeth 
beneath the orbit ; basipterygoid process absent. Otophysic connection, if 
developed, never with an intimate connection with the ear. Palatine formed by 
distinct autopalatine (when ossified) and dermopalatine. Circumorbital series with 
supraorbital, antorbital and six infraorbitals. Rostral ossicles developed. Func- 
tional part of the upper jaw formed by the premaxilla and maxilla ; supramaxillae 
present but not capable of independent movement. Mandible with a fossa on the 
inner surface receiving the A w division of the adductor mandibulae ; retroarticular 
absent. Gill arches with separately ossified dermal tooth plates and endochondral 
supports. First suprapharyngobranchial ossified, second suprapharyngobranchial, if 
present, cartilaginous. Urohyal shallow. Opercular series complete. Pectoral 
girdle attached to the neurocranium ; mesocoracoid arch, two or three postcleithra 
and a pectoral splint present. Pelvic splint present. Vertebral column with auto- 
genous neural arches, haemal arches and parapophyses. Epineural and epipleural 
intermuscular bones present. Supraneurals forming a complete series between the 
occiput and the dorsal fin. Caudal skeleton with two free ural centra, the first sup- 
porting two autogenous hypurals, the second supporting four or five upper hypurals. 
Uroneurals free from centra. Nineteen principal caudal fin-rays, 17 of which are 
branched. Scales cycloid, overlapping and marked anteriorly by prominent radii 



FOSSIL AND RECENT 



producing a scalloped margin. Bone cells present within the anterior and lateral 
fields of the scales. Enlarged axillary scales present. 



Suborder ELOPOIDEI Jordan, 1923 

Diagnosis. Elopiform fishes in which the cranium is never broad. Sensory 
canals of the head largely enclosed by bone ; supraorbital sensory canal running 
through parietal ; ethmoid commissure complete, running through dermethmoid. 
Post-temporal fossae large, directed anteriorly ; sub-epiotic fossae absent. Inter- 
calar large, extending anteriorly to contact the prootic, forming a prootic-intercalar 
bridge. First vertebral centrum articulating with facets upon the exoccipitals and 
basioccipital and forming a functional part of the neurocranium. Lateral ethmoid 
not in contact with the parasphenoid ; parasphenoid narrow. Hyomandibular- 
metapterygoid cup developed. Premaxilla small, never associated with a sensory 
canal. Two supramaxillae. Articular (angular) and endosteal articular (articular) 
separate ossifications. Dentition of small villiform teeth on the dermal jaws, 
vomer, parasphenoid, dermopalatine, endopterygoid, ectopterygoidandbasibranchial 
and basihyal dermal plates. Gular plate large, always horizontal. Gill-rakers 
prominent upon the first three gill arches. Supratemporal large, meeting its partner 
in the dorsal mid-line. Caudal fin with seven hypurals ; the bases of the inner fin- 
rays expanded. Scales with concentric circuli in the anterior and lateral fields. 



Family ELOPIDAE Bonaparte, 1846 

Diagnosis (emended). Elopoid fishes in which the body is rounded. Cranium 
shallow with the mouth terminal. Neurocranium in which the roof is flat ; the parie- 
tals retain evidence of middle pit-lines ; the autosphenotic spine and the epiotic 
process are weakly developed ; the post-temporal fossae extend forward to the level 
of the autosphenotics and remain separate from one another ; the dilatator fossa is 
roofed. Otophysic connection absent. Two rostral ossicles present (only known 
in Elops) . Ouadrate/mandibular articulation at or behind the level of the posterior 
orbital margin. Maxilla shallow, extending posteriorly beyond the level of the eye. 
Mandible shallow with a weakly developed coronoid process situated posteriorly. 
Pseudobranchiae present in Recent genus. Dorsal and anal fins short based. Pelvic 
fins originating beneath or behind the dorsal fin. First anal pterygiophore short. 
Urodermal present. Scales with at least six anterior radii, otherwise marked only 
by fine circuli. Lateral line tubes unbranched. 



Genus ELOPS Linnaeus, 1766 

Diagnosis (emended). Elopid fish in which the cranial bones are devoid of 
ornamentation. Dermethmoid without ventro-lateral projections. Circumorbital 
series with a large first infraorbital which meets the supraorbital and so excludes the 
antorbital from the orbital margin ; infraorbital canal without bone-enclosed 



8 ELOPIFORM FISHES 

branches. Union of infraorbital, antorbital and ethmoid commissure sensory canals 
lying within skin. Supramaxillae narrow, posterior member without strengthening 
ridge. Quadrate with anterior and posterior margins of equal length. Ceratohyal 
imperforate. Preoperculum with a slight posterior expansion, the contained sensory 
canal running at the anterior margin of the bone and opening to the surface by 
numerous pores. Vertebral count high, averaging 75. Caudal fin without fringing 
fulcra. 

Type-species. Elops saurus Linnaeus, 1766. 

Remarks. Although many references have been made to Elops in discussions 
of teleostean phylogeny, no complete osteological description exists. The direct 
relevance of such a study to a work of this nature justifies the inclusion of a descrip- 
tion here. Elops hawaiensis Regan has been chosen as the species to be described 
since the size range of available individuals allows an evaluation of ontogenetic 
changes. 

Regan (1909) recognized seven species of Elops ; E. saurus Linnaeus from the 
east coast of Middle America and E. affinis Regan from the west coast, E. sene- 
galensis Regan and E. lacerta Cuvier & Valenciennes from the west coast of Africa, 
E. machnata Forskal, E. hawaiensis Regan and E. australis Regan from the Indo- 
Pacific. The differences between these species are primarily differences in gill-raker, 
vertebral and scale counts. Whitehead (1962) did not consider E. australis a valid 
species and placed it in synonymy with E. hawaiensis. E. machnata is also very 
much like E. hawaiensis, differing only in vertebral counts (63-64 in the former 
against 66-69 in the latter) and by the fact that when the mouth is closed the pre- 
maxillary tooth band is exposed in E. hawaiensis but is covered by the lower jaw 
in E. machnata (Regan 1909 ; Whitehead 1962). Both of these differences appear 
to be trivial. The former is particularly unsatisfactory while specimens of both Indo- 
Pacific species show variable conditions of the ' exposed ' or ' covered ' state. It is 
likely that only one Indo-Pacific morphospecies exists. 

The differences between E. hawaiensis and the type-species are slight and have 
been illustrated by Whitehead (1962). Certain portions of the osteology of Elops 
have been satisfactorily described in recent literature and where this has been done 
the relevant references are cited without further comment. 



Elops hawaiensis Regan, 1909 
(Text-figs. 1-11) 
1909 Elops hawaiensis Regan : 37. 

Diagnosis. See Regan [op. cit.). 

Habitat. Coastal waters of the Indo-Pacific. 

Description. Neurocranium. The neurocranium is three times as long as 
deep. The maximum width of the cranium occurs at the occiput and is equal to 
one and a half times the depth of the neurocranium. The cranial cavity occupies 
about one-third of the total neurocranial length. 



FOSSIL AND RECENT 9 

Much of the cranial roof is formed by the paired frontals. Each frontal is narrow 
anteriorly but widens at the level of the lateral ethmoid and again above the auto- 
sphenotic spine. Above the orbit the lateral margins of the frontals are parallel, 
a resemblance to leptolepids rather than pholidophorids. Posteriorly each frontal 
ends squarely, with a slight overlap above the parietal and pterotic of its side. The 
mutual interfrontal suture is straight in young individuals, where there may be a 
divergence of the bones anteriorly exposing the ethmoid cartilage. In the adult, 
however, the frontals meet one another throughout their length and the suture line 
becomes wavy posteriorly. Ornamentation upon the frontal is confined to the ridges 
associated with the sensory canals and is described below. 

The anterior end of the cranial roof is formed by the median dermethmoid (rostral 
of Gardiner 1963 and Nybelin 1956, 1967a). The dermethmoid is long, unlike that 
bone in pholidophorids and Liassic leptolepids, and overlaps the frontals, a fact 
which reinforces its interpretation as a dermal element. The ethmoid commissure 
runs transversely across the dermethmoid at the level of the maximum width of the 
latter element. The commissure, which is relatively large in young individuals, 
is contained within a bony tube perforated dorsally by two pores. In young in- 
dividuals this bony tube appears to be an ossification distinct from the underlying 
dermethmoid but in older individuals it appears as an integral part of that bone. 

The nasals, although not strictly part of the neurocranium, may be considered 
here as they are closely associated with the roofing bones. Each nasal is tube-like, 
barely larger than the sensory canal that it carries. The nasal is pierced dorsally by 
four large pores. Gosline (1965) reported the presence of a small prenasal ossicle 
in Elops saurus, which according to Nybelin (1967a) may become incorporated in 
the lateral rostral ossicle. I have found no such prenasal in E. hawaiensis. How- 
ever, the possibility of there being a prenasal ossicle in some species of Elops is 
interesting since it may be indicative of a trend among elopiform teleosts, that is, a 
general tendency toward fragmentation of the dermal bones of the snout. 

The parietal is rectangular and meets its partner in the mid-line except posteriorly 
where the supraoccipital intervenes for a very short distance. There are recognizable 
differences in the dimensions of the exposed part of the parietal between young and 
old individuals. In the young the length of each parietal exceeds its width whereas 
in the adult the converse is true. The change in shape is due to two factors : 
differential growth of the parietal in which lateral growth proceeds at a greater 
rate than longitudinal growth, and secondly the frontals overlap the parietals 
to a greater extent in older fish. Ornamentation upon the parietals, like that of 
other cranial bones, is restricted to ridges associated with the sensory canals. 

The postero-lateral region of the neurocranial roof is formed by the dermal portion 
of the pterotic. The dermal portion of the pterotic forms much of the roof of the 
post-temporal fossa. The lateral and posterior faces of the pterotic are formed by 
the thicker but less dense endochondral bone. As is usual among teleosts there is 
no clear demarcation between the dermal and endochondral portions of the pterotic 
(Gosline 1969 reports the presence of separate autopterotic and dermopterotic in 
Alepocephalus) . The lateral face of the pterotic forms the posterior region of the 
dilatator fossa, much of the hyomandibular facet and the roof of the deep 



ELOPIFORM FISHES 



10mm- 



asp 



e com 



so.sc 



pto 




ot sc 



soc 



Fig. i. Elops hawaiensis Regan. Neurocranium in dorsal view ; this and 
Text-figs. 2-5 are based on a specimen of 208 mm S.L. 



epo. pr 



FOSSIL AND RECENT 

soc 



ptf 



ex 




ptc 



5mm 
Fig. 2. Elops hawaiensis Regan. Neurocranium in posterior view. 



subtemporal fossa. A swelling forming an arc above the subtemporal fossa and 
beneath the hyomandibular facet indicates the path taken by the horizontal semicircu- 
lar canal through the pterotic. Anteriorly and posteriorly the semicircular canal 
descends through the prootic and exoccipital respectively before entering the cranial 
cavity. Medially the pterotic forms the lateral wall of the post-temporal fossa. 
The posterior face of the pterotic contacts the epiotic above and the exoccipital 
below the opening of the post-temporal fossa. The union of the pterotic and 
exoccipital may only be seen in posterior view if the cap-like intercalar is removed. 

The epiotic is seen in dorsal and posterior views of the neurocranium. From the 
former aspect it is represented by a knob-like process, protruding posteriorly from 
beneath the parietal and pterotic. In posterior view the epiotic forms part of the 
dorsal and medial margin of the opening leading to the post-temporal fossa. 

Between the epiotics lies the supraoccipital which, in posterior view, is pentagonal. 
The flat surface of the supraoccipital is interrupted only by a weakly defined vertical 
crest which becomes more prominent dorsally. In dorsal view the supraoccipital 
is seen as a small triangle, the apex of which lies between the posterior ends of the 
parietals. Beneath the roof the supraoccipital extends forward as a median and 
two antero-lateral projections. The latter reach forward to meet the frontal of either 
side and in so doing form the median walls of the post-temporal fossae. The supra- 
occipital is pierced by canals which carry the posterior vertical semicircular canal of 
either side. Laterally, this semicircular canal descends through the epiotic where its 
passage is visible externally as a ridge along the medial margin of the post-temporal 
fossa opening. Ventrally, the posterior vertical semicircular canal runs through the 



12 ELOPIFORM FISHES 

exoccipital before opening to the endocranial cavity immediately above the vagus 
foramen. Ridewood (1904 : 39) noted that in Elops saurus the internal limit of 
the deep subtemporal fossa is formed by the supraoccipital. The supraoccipital of 
E. hawaiensis is not as extensive as that in the type-species and does not play any 
part in the subtemporal fossa. 

The exoccipital exhibits both a lateral and posterior face. The latter face contacts 
the epiotic and supraoccipital above, the basioccipital below, and with its fellow of 
the opposite side surrounds the foramen magnum. In young individuals the margin 
of the foramen magnum is often incomplete ventrally. In its dorsal extent the 
lateral face of the exoccipital is turned sharply inwards to form the posterior wall of 
the subtemporal fossa, while ventrally the exoccipital is inflated and together with 
the prootic and basioccipital forms the outer wall of the saccular recess. The 
saccular recess and the contained otolith are relatively larger in younger individuals. 
Three, occasionally four, foramina pierce the lateral face of the exoccipital. Post- 
eriorly there is a large vagus foramen directed ventro-laterally and slightly 
posteriorly. Anteriorly, close to the exoccipital-prootic suture, there is an antero- 
ventrally directed glossopharyngeal foramen. The third constantly occurring 
foramen is situated immediately in front of the vagal foramen and marks the point 
at which a small branch of the glossopharyngeal nerve left the cranial cavity. This 
accessory glossopharyngeal foramen is, in some specimens, preceded by another 
very small foramen whose function could not be determined ; it may have carried 
a small blood vessel. 

The intercalar ossifies very early in the development of the neurocranium. The 
main body of the intercalar forms a cap over the triradiate union of exoccipital, 
epiotic and pterotic. The ventral limb of the post-temporal is attached to this 
main body. In lateral view the intercalar is produced ventrally and anteriorly. 
The ventral extension partially surrounds the vagus foramen while the well-developed 
anterior extension reaches forward to interdigitate with a posterior outgrowth of the 
prootic. Together the prootic and intercalar form a bridge which stands clear of 
the neurocranial wall, leaving a small foramen which may be seen in ventral view 
(cf. Leptolepis dubia figured by Patterson 1967a : fig. 5). A small branch of the 
glossopharyngeal nerve passes up through this foramen but it is doubtful if this is 
the primary function of the foramen since it is inordinately large for the nerve which 
passes through. 

The basioccipital is ' W '-shaped in cross section, the lateral wings of the ' W ' 
enclosing the saccular recess of either side while the space beneath the central arms 
represents the posterior myodome. The posterior portion of the basioccipital meets 
the paired exoccipitals above in a triradiate suture, and this posterior face articulates 
with a thin vertebral centrum. This first centrum is firmly united with the cranium, 
so much so that in older individuals it is difficult to recognize as a discrete entity. 
Ventrally this centrum bears parapophyses but no pleural ribs, while dorsally there 
is an autogenous neural arch and spine. In small specimens the centrum is still 
pierced by a notochordal foramen. The occipital condyle of lower teleosts is primi- 
tively formed by the inclusion of a centrum into the neurocranium, but rarely is 
there any associated neural arch. Thus Elops would appear to represent a primitive 



FOSSIL AND RECENT 



13 




u 

<v 
+-* 

+-> 

4-1 

a 

5 ttj 



C* 



3 « 






53 







i 4 ELOPIFORM FISHES 

stage, directly comparable to Tarpon but more advanced than that described for 
Heterotis where according to Greenwood (1968) there is also a pleural rib and epi- 
central bones associated with this centrum. It is generally assumed that the 
neural arches associated with this centrum are lost in most teleosts but the situation 
may be more complex since Gosline (1969) records that in Alepocephalus the neural 
arches have become detached from the centrum and are incorporated into the neuro- 
cranium. Clearly the occipital condyle region of the teleost neurocranium warrants 
closer attention. 

The prootic is the largest endochondral component of the neurocranium and, as 
is usual in teleosts, is complicated in shape. The prootic is disposed in a vertical 
and horizontal plane, the former consisting of transverse and longitudinal faces. The 
horizontal part of the prootic extends medially from the lateral face beneath the level 
of the pars jugularis and meets its fellow in the mid-line, so forming the prootic 
bridge. The anterior margin of the prootic bridge is deeply indented in the mid- 
line, forming the posterior margin of the pituitary foramen. The anterior border 
of this foramen is formed by the basisphenoid. Posteriorly the prootic bridge is 
sutured with the basioccipital so separating the cranial cavity above from the pos- 
terior myodome below. The abducens nerve pierces the prootic bridge near to its 
junction with the lateral vertical face of the prootic. 

The lateral vertical face of the prootic is depressed postero-dorsally where it 
forms the anterior region of the subtemporal fossa. Postero-ventrally the prootic 
is inflated to form part of the wall of the otolith chamber. Within the lateral face 
of the prootic runs the long pars jugularis. Posteriorly the pars jugularis opens to 
the lateral face of the neurocranium by the jugular foramen, close to the prootic- 
exoccipital suture. Immediately above the jugular foramen the prootic is raised 
into a short spine which extends posteriorly and interdigitates with an anterior limb 
of the intercalar. The hyomandibular foramen opens to the surface of the prootic 
above the level of the pars jugularis to which it is connected by a short canal. Above 
the foramen the path taken by the hyomandibular ramus is seen as a faint groove. 
From the floor of the pars jugularis a short canal passes ventrally through the 
prootic and splits into two branches, one opening on the medial face, the other on 
the lateral face. The former carries the palatine branch of the facial, the latter 
carries the orbital artery. 

The medial wall of the pars jugularis is perforated by a single large foramen 
through which pass the roots of the trigeminal, facial and profundus ciliaris nerves. 
The pars ganghonaris is simply represented by a small depression on the medial 
aspect of the prootic. The gasserian and geniculate ganglia are lodged within this 
depression. 

The anterior, transversely orientated face of the prootic is small. It meets the 
autosphenotic and pterosphenoid dorsally, the basisphenoid ventro-medially and 
the parasphenoid ventro-laterally. Medial to the large anterior opening of the pars 
jugularis there is a smaller oculomotor foramen. 

The autosphenotic is tetrahedral, the apex of the tetrahedron being produced 
laterally as a small spine to which the levator arcus palatini musculature is attached. 
Laterally the autosphenotic forms the anterior part of both the dilatator fossa and the 



FOSSIL AND RECENT 

ors 



15 



I so. s.c 



pto 



asp 




Mot 



myp 



Fig. 4. Elops hawaiensis Regan. Neurocranium in orbital view. 



hyomandibular facet. The orbital face of the autosphenotic contacts the prootic, 
pterosphenoid and frontal and is pierced by a foramen conveying the otic branch of 
the facial to innervate the otic sensory canal. Internally the autosphenotic forms 
the anterior wall of the post-temporal fossa. Ridewood (1904) stated that in Elops 
saurus the anterior limit of the post-temporal fossa is formed by the pterosphenoid 
(alisphenoid). This is not the case in E. hawaiensis. 

The lateral margins of the heart-shaped optic foramen are formed by the ptero- 
sphenoid of either side which lies medial to both the prootic and autosphenotic. 
The surface of the pterosphenoid is marked by a groove which curves upwards and 
forwards from the opening to the pars jugularis. This groove contains the super- 
ficial ophthalmic branches of V and VII. The trochlear nerve passes out of the cranial 
cavity through the optic foramen. 

The orbitosphenoid lies anterior to the pterosphenoids. The anterior margin of the 
orbitosphenoid is perforated by a large foramen through which pass the olfactory 
tracts. Ventrally the orbitosphenoid is produced as a membranous interorbital 
septum. Posteriorly the orbitosphenoid forms the dorsal margin of the optic fora- 
men. The ventral margin of the optic foramen is formed by the ' wings ' of the 
' Y '-shaped basisphenoid which contact the prootic and pterosphenoid of either 
side. The laterally flattened stem of the basisphenoid passes antero-ventrally to meet 
the parasphenoid so dividing the entrance to the posterior myodome. The basi- 
sphenoid stem ossifies late in ontogeny. 

The parasphenoid is long, extending from its contact with the vomer anteriorly 
to the hind edge of the basioccipital. Beneath the orbit the parasphenoid is relatively 
narrow but beneath the otic region it is broader as it forms the floor of the posterior 



i6 



ELOPIFORM FISHES 



10 mm 




fahm 



exo 



fuv 



boc 



Fig. 5. Elops hawaiensis Regan. Neurocranium in ventral view. 
Hatched areas represent cartilage. 



FOSSIL AND RECENT 17 

myodome. The parasphenoid is produced dorsally as a shallow ascending wing 
which partially overlaps the prootic. A foramen for the internal carotid artery 
occurs within the parasphenoid at the base of the ascending wing. Posteriorly 
the parasphenoid ends in a notch forming the lower edge of a foramen leading to the 
myodome, the dorsal edge of this foramen being formed by the basioccipital. Beneath 
the orbit the parasphenoid bears teeth. In young individuals the teeth are very 
small and clustered along the margins of the bone but in adults the pattern changes 
to a dense and complete covering of small villiform teeth. The parasphenoid is 
usually pierced in the mid-line beneath the basisphenoid by a foramen for the bucco- 
hypophysial canal but in the specimen figured here (Text-fig. 5) it is closed ventrally. 

The vomer consists of a broad head and a spear-shaped shaft which extends 
posteriorly beneath the lateral ethmoid. The paired vomerine tooth plates are 
raised upon short pedicels. In small individuals each toothed area is separated 
from its partner and bears about 20 small recurved teeth. During growth the 
number of teeth increases to about 60 and the toothed areas of either side approxi- 
mate to one another in the mid-line. Even in the largest specimens examined the 
toothed area retains its identity as a paired element. 

Much of the snout region is formed by cartilage. Between the dermethmoid 
above and the vomer below the ethmoid cartilage separates the nasal capsules 
of either side. The posterior, ventral and anterior walls of the nasal capsule are 
composed of ethmoid cartilage. Beneath the anterior region of the solum nasi the 
cartilage is somewhat denser than elsewhere and forms an oval facet for the articula- 
tion of the palatine head. Two facets for jaw articulation occur upon the lateral 
face of the cartilage above the vomer. The posterior of these receives the head of 
the maxilla while the anterior receives the premaxilla. In the largest individuals 
the extreme anterior tip of the ethmoid cartilage may show a limited amount of 
endochondral ossification. Posteriorly the lateral edge of the planum antorbitale 
ossifies perichondrally to form the lateral ethmoid. The olfactory tract and the 
nasal branch of the orbitonasal artery pierce the planum antorbitale through a 
large foramen olfactorium advehens. There is no anterior myodome. 

Hyopalatine bones. The hyomandibular is a stout element with a single broad 
articulatory head (a distinction from the type-species which shows a weakly divided 
head, Ridewood 1904) which is inclined antero-ventrally. The hyomandibular is 
produced postero- ventrally as a narrow shaft. Dorsally the posterior margin of 
the hyomandibular projects as a prominent opercular process which articulates with 
a cup-shaped depression on the operculum. The hyomandibular ramus of VII and 
the efferent hyoidean artery pierce the hyomandibular immediately beneath the 
anterior limit of the head and run obliquely through the bone to emerge beneath 
the opercular process before running down in a groove on the posterior edge of the 
shaft. Anteriorly the head of the hyomandibular is produced as a thin wing of 
bone which lies medial to the dorsal part of the metapterygoid. The cup-shaped 
space left between the hyomandibular and metapterygoid, the hyomandibular- 
metapterygoid cup, receives deeper fibres of the levator arcus palatini. 

The metapterygoid is irregular in shape. Its dorsal margin is turned horizontally, 
medial to that part referred to above which overlaps the hyomandibular. The 



18 ELOPIFORM FISHES 

anterior border of the metapterygoid overlaps the endopterygoid, while the ventral 
margin is in synchondral union with the quadrate. 

The quadrate bears a stout articulatory condyle which in adult fish lies vertically 
beneath the head of the hyomandibular. The posterior margin of the quadrate 
is thickened. Upon the medial surface of the quadrate a deep groove runs close 
to the posterior margin and accommodates the ventral end of the stylif orm symplectic. 
The dorsal end of the latter is united with the hyomandibular shaft by cartilage. 

The ectopterygoid is composed of two limbs, the posterior grooved to fit the 
anterior margin of the quadrate and the longer anterior limb disposed more horizon- 
tally. The dorsal edge of the ectopterygoid is thickened and to this thickening is 
attached a tough fold of skin which inserts on the medial face of the overlying infra- 
orbitals. Gosline (1965) pointed out that there was no ectopterygoid process in 
Elops but the thickened ridge in E. hawaiensis may be considered as such a rudi- 
mentary process. The oral surface of the ectopterygoid bears many small villiform 
teeth. 

The convex oral surface of the endopterygoid is covered with many small villiform 
teeth, continuous with those upon the ectopterygoid. 

There is a small dermal palatine (dermopalatine) but no ossified autopalatine. 
The palatine cartilage overlaps the anterior ends of the ectopterygoid and endoptery- 
goid posteriorly. Anteriorly the cartilage is slightly swollen and bears a facet upon 
the dorso-medial surface which fits against the floor of the nasal capsule. The 
ventro-lateral surface of the palatine cartilage articulates with the maxilla but there 
is no special facet developed. The dermopalatine is represented by a tooth plate. 
The teeth borne by this plate are continuous with those on the ectopterygoid and 
endopterygoid but are more like those of the vomer in shape, being pointed and 
recurved. 

Dermal upper jaw. The upper jaw extends from the tip of the snout to behind the 
orbit. The relative length of the upper jaw increases slightly throughout life. The 
premaxilla forms a little less than one-third of the convex oral margin. The pre- 
maxilla is relatively shallow, the maximum depth being equal to less than one-quarter 
of its length and situated towards the anterior end. From its point of maximum 
depth the bone tapers in both directions. The oral surface of the premaxilla bears a 
band of small teeth. Each tooth is conical and bears a pointed enamel cap, set at an 
angle to the main axis of the tooth. Basally the tooth is set within a shallow socket. 

The maxilla is elongate and deepest posteriorly. Anteriorly the maxilla curves 
both dorsally and medially to end in a simple rounded head which articulates directly 
with the ethmoid. Behind the maxillary head the dorsal surface of the maxilla 
bears a palatine process which articulates with the ventro-lateral aspect of the 
palatine through a biconcave sliver of cartilage. The oral margin of the maxilla 
bears a band of small teeth similar to those upon the premaxilla. Posteriorly 
the tooth band extends on to the lateral surface of the bone. 

There are two supramaxillae. The anterior supramaxilla is produced anteriorly 
as a spine. The posterior bone also bears an anterior spine which overlies part of 
the anterior supramaxilla. Both supramaxillae are without strengthening ridges 
and are relatively immobile, unlike these bones in clupeoids. 



FOSSIL AND RECENT 



19 




1_ 




1_ 




J3 


i 







<_> 


'> 


i/i 




d 







0) 


Q. 










+-> 




VH 




<D 




■— 1 




.a s 




<u 




6 S 




.2 "3 




jH <U 




rt cn 




u <n 




O -2 




ni 




I-H 




. 0) 




a > 




ni <o 




be <" 




<U <+H 




« 




to £ 

<2 en 




S O 




•~ 9? 




1 s 

8 




« 




to 




£ 



UJ 






20 ELOPIFORM FISHES 

Mandible. Little may usefully be added to what Ridevvood (1904) has described 
in Elops saurus. Mention may be made of the deep fossa on the medial surface of 
the dentary which receives the A w division of the adductor mandibulae. This fossa 
is relatively deep in Elops, its development probably being advanced and perhaps 
correlated with the predatory habits of this genus since it is absent in the micro- 
phagous clupeoids. 

Circumorbital series. The circumorbital ring is incomplete dorsally. The supra- 
orbital is the only anamestic component of the series. It is a narrow bone which lies 
close against the frontal. The ventral end of the supraorbital is overlain by the 
antorbital. This latter bone and the snout ossicles have been described and figured by 
Nybelin (1956, 1967a) in Elops saurus and the situation is the same in E. hawaiensis. 

The first infraorbital is very deep anteriorly but shallow posteriorly. Charac- 
teristically the first infraorbital contacts the supraorbital and so excludes the antor- 
bital from the orbit. The second infraorbital is very short and shallow and meets 
the expanded third infraorbital in an oblique suture. The fourth and fifth infra- 
orbitals, like the third, are longer than deep. The sixth infraorbital, or dermo- 
sphenotic, consists of a shallow base which is produced antero-dorsally as a process 
which lies tightly against the frontal. 

Hyoid arch, gill arches and gular plate. These structures and associated dermal 
elements have been dealt with in various papers ; the hyoid and gill arches by 
Ridewood (1904) and especially Nelson (1968a, b, E. hawaiensis, and 1969a), the 
dentition of the mouth cavity by Nybelin (1968) and the gular plate by Jessen 
(1968). Although the above authors described Elops saurus there is little difference 
in E. hawaiensis. 



dhh 




vhh ace 



3mm 

1 



Fig. 7. Elops hawaiensis Regan. Hyoid bar and branchiostegal rays in left lateral view ; 
this and Text-figs. 8-n based on an alizarin preparation of B.M.N.H. 1962. 4. 3.1, 
62 mm S.L. 



Tp bh 



Tpbb 1+2*3 



hb3 



Fpbb4 



FOSSIL AND RECENT 

bh 



21 




sb1 



Tp ib 4 



3 mm 



Fig. 8. Elops hawaiensis Regan. Gill arches of right side in dorsal view. The dorsal 
elements have been turned back exposing their ventral aspect ; the second supra- 
pharyngobranchial is obscured by ib3, and the fourth infrapharyngobranchial is covered 
by Tpib4. Hatched area represents cartilage, black represents tooth plates and gill-rakers. 



The only additional feature worthy of note is the attachment of the gill arches to 
the neurocranium, an attachment which is effected through the first infrapharyngo- 
branchial and suprapharyngobranchial. The former element is attached through 
a small disc of cartilage to the parasphenoid behind the internal carotid foramen 
in that bone. A posterior attachment to the neurocranium is effected by a ligament 
which passes from the dorsal limb of the first epibranchial to the anterior extension 
of the intercalar. Within this ligament the first suprapharyngobranchial is repre- 
sented as a small ossified rod situated just above the first epibranchial. 



22 ELOPIFORM FISHES 

Cephalic sensory canal system. The main canals run within bony tubes which are 
perforated by pores. ' The sensory canals produce ridges upon the bones through 
which they run. The junctions between the major canals lie in the skin. 

The lateral line passes on to the cranial roof by a pore at the postero-lateral 
corner of the pterotic. The otic canal runs through the extreme lateral edge of the 
roofing portion of the pterotic and opens anteriorly above the autosphenotic spine. 
The dorsal surface of the otic canal is perforated by three or four pores. The pos- 
terior opening of the otic canal lies adjacent to a pore in the supratemporal and also 
adjacent to but at a distance from the top of the preopercular canal. The triradiate 
union between the supratemporal commissure, otic and preopercular canals occurs 
within the skin. 

The anterior opening of the otic canal lies adjacent to two pores ; one within the 
frontal which is the lateral opening of the supraorbital canal and one from the dorsal 
surface of the dermosphenotic which represents the dorsal limit of the infraorbital 
canal. Again, the union of these canals lies in the skin. 

From its union with the otic and infraorbital canals, the supraorbital canal curves 
medially and then anteriorly to run the length of the frontal. A large branch of 
the supraorbital canal is given off posteriorly and this runs back on to the parietal 
where it opens to the surface by a single terminal pore, behind which there is in some 
specimens a continuing shallow groove. The groove is interpreted as representing 
the anterior pit-line. The parietal is also marked by three small pores arranged in a 
transverse row. Such a pore arrangement is similar to the middle pit-line of hale- 
costomes and leptolepids. In these latter fishes the pores are set in a shallow groove 
but no such connecting groove exists in Elops. The main supraorbital sensory canal 
opens to the surface by several pores on both the lateral and medial aspects of the 
canal. Above the epiphyseal region there is a predominance of medial pores, while 
above the orbit and otic regions lateral pores are more common. 

The infraorbital canal runs very close to the orbital margin of the infraorbitals, 
opening to the surface by a series of posterior and inferior pores. The neuromast 
distribution has been noted by Nelson (1969b). Anteriorly the infraorbital canal 
joins with the antorbital branch and the ethmoid commissure. The distribution 
of the sensory canals in this region has been described by Nybelin for Elops lacerta 
(1956, 1967a) which in all respects is similar to E. hawaiensis. 

The preopercular canal runs at the anterior margin of that bone and in its ventral 
half opens to the surface by ten to fifteen pores. Ventrally the preopercular canal 
is continuous with the mandibular canal which is contained within the articular and 
dentary. The mandibular canal opens to the surface by a series of large pores. 

Opercular series. The opercular bones of Elops saurus have been described by 
Ridewood (1904). In most respects they are similar to those of E. hawaiensis, 
differing only in proportions. Further description is unnecessary. 

Pectoral girdle and fin. The supratemporal, although not strictly part of the 
shoulder girdle, may be considered here. As in all elopoids the supratemporal is 
very large and thin, fully justifying its alternative name of the ' scale bone '. The 
supratemporal has a straight anterior margin which extends from the postero-lateral 
corner of the pterotic to the mid-line. The posterior margin is broadly curved, 



FOSSIL AND RECENT 



23 



sea 




pc 1 1-3 



r p1 



cor 



3mm 



Fig. 9. Elops hawaiensis Regan. Pectoral girdle of the left side in medial view. 



showing an indentation in the middle of its length. The growth pattern of the 
supratemporal refutes any suggestion that it is formed of two elements. The 
supratemporals of either side meet one another in the mid-line covering both the 
epiotics and the supraoccipital. In consequence of this median union the supra- 
temporal commissure is completely enclosed by bone. The latter canal opens to 
the surface by a few posterior pores. The main lateral line pierces the ventral surface 
of the supratemporal. 

The post-temporal is composed of a main body which bears three processes. The 
body of the bone is a small flat plate on the underside of which there is a small 
depression which receives the dorsal end of the supracleithrum. The lateral line 
runs through the lateral margin of the post-temporal. Two of the three processes 



24 ELOPIFORM FISHES 

mentioned above are merely extensions of the anterior margin of the main body. 
Medially there is a pointed dorsal limb which overlies the epiotic process. Laterally, 
there is a very small limb which reaches forwards but fails to reach the pterotic. 
The third process originates from the undersurface of the main body and passes 
antero-ventrally to become firmly anchored to the intercalar. 

The supracleithrum is long, narrow and inclined postero-ventrally. The anterior 
margin is slightly thickened. The lateral line runs obliquely through the upper half. 
Ventrally the supracleithrum overlaps the cleithrum. The latter element is, as 
usual, the largest single component of the shoulder girdle. From a narrow dorsal 
portion the bone expands as it curves antero-ventrally and medially to contact its 
fellow in the ventral mid-line. The anterior margin of the cleithrum is turned 
inwards to form both the posterior wall of the gill chamber and a site of origin for 
the sterno-hyoideus musculature. There are three narrow postcleithra attached to 
the medial surface of the cleithrum. 

The endochondral coracoid is attached to the ventral limb of the cleithrum. In 
its middle region it fails to contact the cleithrum and there is instead a large space, 
the interosseous foramen of Starks (1930). Posteriorly the coracoid contacts the 
scapula laterally and the mesocoracoid dorsally. 

The scapula spreads over the inner face of the cleithrum. The scapula completely 
encloses the scapular foramen. Posteriorly the margin of the scapula bears a deep 
notch, the point of insertion of the outermost fin-ray. The mesocoracoid is repre- 
sented by a thin bar of bone which passes dorsally from its union with the scapula 
and coracoid. The dorsal part of the mesocoracoid is expanded and like the scapula 
spreads over the inner face of the cleithrum. 

Support of the dermal fin-rays is shared by the endochondral girdle and the radial 
elements. The outer ray is attached directly to the scapula. There are four ossified 
proximal radials which articulate with the coracoid. The longest of the proximal 
radials is the innermost, the shortest the outermost. A further row of cartilaginous 
distal radials may be recognized but neither their number (usually six to eight) nor 
their distribution is constant from individual to individual or even from one side of a 
specimen to the other. The distal radials appear to play little part in the support 
of the fin. 

There are 15-16 branched rays plus the outer unbranched ray. The longest is 
the second in the series, the outermost branched ray. At the base of the first ray 
there is a small triangular pectoral splint bone which is positioned equally over 
both halves of the ray. 

Pelvic girdle and fin. The pelvic fin originates beneath the anterior half of the 
dorsal fin. The fin-rays are supported by both radials and a pelvic bone. The 
latter is triangular, the apex directed anteriorly. Posteriorly the margin of the pelvic 
bone is covered by cartilage. Two or three cartilaginous radials articulate with the 
posterior margin. An ossified inner radial articulates with the postero-medial aspect 
of the pelvic bone. This inner radial is represented by a curved splint, the proximal 
end of which lies between the bases of the inner four rays. 

There are 15-16 pelvic fin-rays of which the outermost is unbranched. The third 
ray is the longest of the series. At the base of the upper half of the outermost ray 



FOSSIL AND RECENT 25 

there is an elongate pelvic splint bone. The anterior end of this pelvic splint lies 
free in the somatic musculature. 

Vertebral column. The vertebral column is composed of 68 vertebrae. There 
are 19 caudal vertebrae. 

Each centrum is amphicoelous and pierced centrally for the passage of the noto- 
chord. The majority of the centra are as deep as long but those anteriorly are 
deeper than long while those posteriorly are slightly longer than deep. The centra 
are marked laterally by fine grooves separated by equally fine ridges. The neural 
arches, haemal arches and parapophyses are received within shallow pits upon the 
centra. 

Anteriorly the neural arches are wide and expanded distally. Towards the pos- 
terior end of the abdominal region the neural arches become more slender and there 
is no distal expansion. The neural spines of the first 34 vertebrae remain as separate 
halves but posterior to this level the lateral halves fuse to form solid median spines. 
The neural spines of the third, fourth and fifth preural centra are somewhat longer 
and straighter than those immediately in front and aid in the support of the caudal 
fin-rays. 

The first 47 centra bear parapophyses. To these are attached the pleural ribs. 
The anterior parapophyses are small but those associated with centra 35-47 gradually 
increase in length. Behind this level complete haemal arches and spines are seen, 
the last four being long and stout and involved in the support of the caudal fin-rays. 

Epineural intermuscular bones are only absent from the centrum incorporated 
into the neurocranium and the last three caudal vertebrae. The epineurals asso- 
ciated with the first 30 or so vertebrae are attached to the bases of the neural arches. 
Those behind lie free, the anterior free members having forked bases. Epipleural 
intermusculars are found throughout the caudal region. All lie free, and like their 
neural counterparts those found anteriorly show forked bases. 

Sigmoid supraneurals occur between the tips of the first 32 neural spines. The 
first supraneural is the largest, thereafter they decrease in size posteriorly. It is 
of interest to note that supraneurals occur between the first few pterygiophores of 
the dorsal fin. 

Median fins. The dorsal fin lies in the centre of the back and consists of 25 
fin-rays of which the first four, although paired, are unbranched and not articulated. 
The seventh ray is both the longest and the first branched of the series. The fin- 
rays are supported upon 21 pterygiophores of which all except the first two consist 
of distinct proximal, middle and distal radials. The distal radials lie between the 
bases of the fin-rays. The first two pterygiophores consist of two radial elements, 
fusion having occurred between the proximal and middle radial. Together, the first 
two pterygiophores support the first five fin-rays. 

The origin of the anal fin is midway between the pelvic fin and the caudal peduncle. 
The anal fin is composed of 17 fin-rays of which the first three are not articulated. 
The sixth ray is both the longest and the first branched of the series. Thirteen 
pterygiophores support the fin-rays. The first is long and like its dorsal counterpart 
consists of fused proximal and median radials and a separate distal radial. The 
first two pterygiophores support the first six fin-rays. 



26 



CSC 



ELOPIFORM FISHES 

ep1-3 un1-3 



nsp 



pu4 



2 mm 




Fig. io. 



Elops hawaiensis Regan. Caudal skeleton in left lateral view. 
Arrows indicate upper and lower principal fin-rays. 



The caudal fin skeleton of Elops has been figured and described several times, 
the most important references being those of Regan (1910), Hollister (1936) and 
Nybelin (1963, 1971) and although those works were concerned with Elops saurus, 
E. lacerta and E. senegalensis , the caudal structure of E. hawaiensis is very similar 
(Text-fig. 10). 

Squamation. There are 95-97 scales in the lateral line series. The transverse 
count immediately in front of the dorsal fin is 12 above and 17 below the lateral line 
series. The scales are relatively small and approximately circular in shape. The 
anterior margin tends to be straight. The nucleus is positioned centrally and the 
surface of the scale is marked by concentric circuli which fade out on the exposed 
surface. The posterior margin of the scale is often frayed during life giving the 
margin a crenulated appearance. The anterior field is marked by 9-13 strong radii 
which interrupt the anterior margin producing a scalloped edge. The scales from 
the posterior region of the body generally show fewer radii. Bone cells are found 
within the scales and are most abundant in the anterior fields. 

Lateral line scales are distinguished by the presence of a prominent tube con- 
taining the lateral line canal. At the bases of the fins there are many smaller scales 



FOSSIL AND RECENT 



27 









a 
W 



E 
E 
o 



c 






o 

l-H 



28 ELOPIFORM FISHES 

which tend to be ovoid and to lack radii. Together these small scales form a basal 
sheath to the dorsal and anal fins. Enlarged axillary scales are present above the 
paired fins. 



Appendix to the genus Elops 

Woodward (1901) described a fossil form which he referred to the genus Elops 
without any distinct species recognition. The Elops sp. is represented by two 
specimens (B.M.N.H. 39443 and P. 1762) from the London Clay (Ypresian) of Shep- 
pey, Kent, England. Unfortunately the specimens are poorly preserved and have 
suffered from pyritization so that very little may usefully be added to Woodward's 
(1901) description. Elops sp. is certainly larger than any extant species and shows 
a depression upon the skull roof. This latter feature, although obvious, is of little 
use taxonomically since the largest individuals of the extant species also show a 
depression of limited size. The scales of the Eocene Elops are marked in their 
exposed portion by ridges, but again this is sometimes seen in the Recent species. 
All features that may be seen of the cranial anatomy, vertebral column and squama- 
tion suggest that this species is correctly referred to the genus Elops but lack of 
material results in the Eocene form being left as Elops sp. 



Genus DAVICHTHYS nov. 

Diagnosis. Elopid fishes reaching 120 mm S.L. Head shallow, the length of 
the cranium forming about 30 per cent S.L. Cranial bones generally smooth. 
The first infraorbital is rounded and fails to contact the supraorbital. Infraorbital 
canal with a few moderately long branches ; the union of the antorbital and infra- 
orbital canals not enclosed by bone. Maxilla extending to the hind margin of the 
orbit ; two supramaxillae, the posterior marked by a pronounced strengthening 
ridge. Quadrate with the anterior margin considerably longer than the posterior 
margin ; quadrate/mandibular articulation well behind the eye. Ceratohyal solid. 
Preoperculum narrow throughout, the contained sensory canal opening directly 
to the surface by a few large pores. Vertebral column with 50-56 vertebrae of 
which 20 are caudal. Pectoral and pelvic fins each composed of 12 rays. Scales 
small, circular and those in the anterior region of the body marked by four or five 
weakly defined anterior radii. 

Remarks. The genus is founded on the best known species, Davichthys dubius 
(Davis) which had been questionably placed in synonymy with Osmeroides gracilis 
Davis, to which it shows little similarity. 0. gracilis is made the type-species of a 
new albulid genus Lebonichthys on p. 171. Davichthys is known from the Lower 
Cenomanian of Morocco and from the Middle Cenomanian and the Upper Santonian 
of Lebanon. 

Type species. Osmeroides dubius Davis. 



FOSSIL AND RECENT 29 

Davichthys dubius (Davis) 
(Text-figs. 12, 13) 

1887 Osmeroides dubius Davis : 565, pi. 31, fig. 4. 
1901 lOsmeroides gracilis Davis ; Woodward : 16. 

Diagnosis. Davichthys reaching 70 mm S.L. Skull roof with faint ornamentation 
on the frontal bones above the orbit. Maxilla with straight oral margin ; supra- 
maxillae large ; premaxilla shallow, forming about one-quarter of the oral margin 
of the jaw. Preopercular sensory canal running at the anterior margin of the bone. 
Dorsal fin situated in the centre of the back and composed of 15 rays. Anal slightly 
nearer to the caudal peduncle than to the pelvic fin and composed of 16-17 rays. 
Caudal fin with fringing fulcra. 

Holotype. Nearly complete fish, R.S.M. 1891.59.38, from the Upper Santonian 
of Sahel Alma, Lebanon. 

Material. The holotype and the following specimens were examined : R.S.M. 
1891.59. 147, B.M.N.H. 48155 and 46539. All are from the Upper Santonian of 
Sahel Alma, Lebanon. 

Description. The anatomy of Davichthys is very like that of Elops and Anae- 
thalion and distinguishing characters are few and minor. In several respects 
Davichthys is intermediate between Elops and Anaethalion. 

The head, as in all elopids, is relatively shallow and the gape moderately large. 
The skull roof is similar to Elops and Anaethalion in being flat, with medially united 
parietals and with frontals which are narrow and parallel-sided above the orbit but 
widen considerably above the otic region. In contrast to Elops and Anaethalion 
the frontal of Davichthys is ornamented above the orbit by weak ridges. The 
dermethmoid is flat, as in all elopoids, and does not bear the ventro-lateral pro- 
jections seen in some species of Anaethalion (Nybelin 1967b : fig. 4B). The absence 
of such projections is a resemblance to Elops. An ethmoid commissure could not 
be identified in the specimens examined. 

The neurocranium is shallow with the narrow, straight parasphenoid lying almost 
horizontally, as in other elopids. Anteriorly the vomer bears a rounded tooth plate 
which carries a dentition of small pointed teeth. Whether the vomerine tooth plate 
was paired or not could not be seen. 

The hyomandibular is attached to the neurocranium by a single broad head and 
the shaft is inclined posteriorly, as in Elops and some species of Anaethalion. While 
most elopids have an undivided hyomandibular head Elops saurus shows a clear 
division into anterior and posterior portions (cf. Ridewood 1904 : fig. 10). 

The metapterygoid overlaps both the hyomandibular and the endopterygoid and 
is unusual among elopids in being deeper than long. The quadrate is also distinctive 
in that the anterior border is longer than the posterior. As in Elops and Anaethalion 
the quadrate does not have a well-defined facet behind the articulatory condyle as 
it does in megalopids. In the remainder of the hyopalatine series there is a close 
similarity with other elopids. Thus the ectopterygoid is narrow and shows a ridge 



3° 



ELOPIFORM FISHES 



5 mm 



pmx 




a pa 



smx 1-2 



ecp art s y 



lop 



ao. sc 




io 5 



i o. s. c 

Fig. 12. Davichthys dubius (Davis). Above, cranium in left lateral view. 
Below, circumorbital bones. Restoration based on several specimens. 

which probably represents a rudimentary ectopterygoid process and the ectoptery- 
goid, endopterygoid and dermopalatine bear many tiny teeth. The autopalatine 
is weakly ossified and the anterior end is produced in a simple rounded knob. 

The upper jaw is like that of other elopids. The premaxilla is narrow compared 
to that bone in Elops and the maxilla, which is deeper than that of Elops, bears an 
ill-defined palatine process. This small palatine process is more like the bony 
nubbin seen in Anaethalion than the well-developed process in Elops. There are 
two supramaxillae, the posterior of which bears a prominent strengthening ridge as 
in Anaethalion. The dentition of the upper jaw consists of a band of tiny pointed 
teeth on both the maxilla and premaxilla. 

The mandible is similar to that of other elopids in being shallow throughout with 
a small coronoid process situated posteriorly and in having the mandibular sensory 



FOSSIL AND RECENT 



3i 




E 






c^ 



00 



o 

<n 

m 



13 
CD 



0) 



a 
W 



Q 



•3 



£ 






32 ELOPIFORM FISHES 

canal wholly contained within bone. The oral margin of the dentary is straight, 
unlike the concave outline of the dentary of Elops. The dentition is similar to that 
of Elops. The quadrate/mandibular articulation occurs beneath the otic region. 

The bones surrounding the orbit are set in typical elopid fashion. The nasals are 
slender tubes which probabby lay free in the skin. The splint-like supraorbital lies 
in tandem with the triangular antorbital. Five infraorbitals were recognized but a 
dermosphenotic was not seen. The first infraorbital is rounded anteriorly as in 
Anaethalion and thus differs from the Elops condition. In shape the posterior 
infraorbitals are very similar to those of other elopids, but in their width there is a 
closer resemblance to Elops. The infraorbital sensory canal runs close to the orbital 
margin and sends off bone-enclosed branches within the second, third and fourth 
infraorbitals. These bone-enclosed branches are a resemblance to Anaethalion 
rather than Elops. Anteriorly the main infraorbital canal passes straight out of the 
bone. The narrow antorbital branch must have linked with the infraorbital canal 
outside the bone, as in Elops but unlike Anaethalion (Nybelin 1967b). 

The anterior and posterior ceratohyals are the same shape as in Elops and extend 
between the jaw rami. As in Elops the anterior ceratohyal is solid, not fenestrated 
as it is in Anaethalion. Little may be seen of the branchial arches, but the first 
and second gill arches bore long gill-rakers, a feature which is totally unlike Lebonich- 
thys ( = Osmeroides) gracilis to which Davichthys dubius had previously been allied. 

The opercular series is very much like that of Anaethalion. The preoperculum is 
narrow throughout, unlike Elops, but like that genus the sensory canal runs at the 
anterior margin in contrast to the condition in Anaethalion. 

The postcranial skeleton is little different from either Elops or Anaethalion. The 
vertebral column consists of 50 vertebrae of which approximately 19 are caudal. 
The total count agrees more closely with Anaethalion (42-66) than with Elops 
(63-80) but in all three genera there are approximately the same number of caudal 
vertebrae. 

The dorsal fin is situated in the middle of the back as in Elops but unlike most 
species of Anaethalion except A. ?cf. subovatus (Nybelin 1967b : pi. 5). 

A distinction from both Elops and Anaethalion is the length of the pectoral 
fin-rays, which extend posteriorly almost to the pelvic fin. 

The caudal skeleton is, as far as may be seen, similar to both Elops and at least 
some species of Anaethalion. In retaining fringing fulcra there is a greater similarity 
with Anaethalion. 

The cycloid scales are small, circular and marked by fine circuli. Those in the 
anterior region of the body have three or four radii in the anterior field which produce 
a scalloped anterior margin. 

Davichthys gardneri sp. nov. 

(Text-figs. 14, 15) 

Diagnosis. Davichthys reaching 120 mm S.L. Skull roof without ornament. 
Maxilla with straight oral border ; supramaxillae slender, the second particularly 
shallow posteriorly. Premaxilla forming nearly one-third of the upper jaw margin 



FOSSIL AND RECENT 

hm 



33 



pmx 




- sop 



5mm 



art 



Fig. 14. Davichthys gardneri sp. nov. Cranium and pectoral girdle in left 
lateral view. Based on the holotype. 



and moderately deep anteriorly. Preopercular canal distant from the anterior 
margin of the bone. Dorsal fin situated mid- way between the occiput and the caudal 
peduncle and composed of 23 rays. Anal situated mid-way between the pelvic fin 
and the caudal peduncle, with 14 rays. Fringing fulcra absent. 

Holotype. Nearly complete fish, B.M.N.H. 49520, from the Middle Cenomanian 
of Hakel, Lebanon. 

Material. No material other than the holotype is known. 

Remarks. This species is founded on a single specimen previously identified as 
Halec. The essential characters of the genus Davichthys are to be found here, the 
differences from the type-species being those meristic counts and morphometric 
ratios which may be inferred from the specific diagnoses. 

Davichthys gardneri is larger than the type-species and has a relatively deeper 
head. The dermethmoid is produced laterally as a small prong resembling the larger 
process in Anaethalion. 

The hyomandibular of D. gardneri shows a very well-developed opercular process 
and the shaft is narrow, differing in both features from D. dubius. The upper jaw 
differs from that in the type-species in showing a prominent palatine process on the 
maxilla, very shallow supramaxillae and a premaxilla which is substantially deeper. 

The preoperculum appears slightly more primitive than in the type-species, 
having no lower limb and the sensory canal distant from the anterior margin. The 



34 



ELOPIFORM FISHES 



nu1 



ep1-3 




un 3 



h 4-7 



Fig. 15. Davichthys gardneri sp. nov. Caudal skeleton in left lateral view. 

Based on holotype. 

operculum is narrower than in D. dubius and the ventral margin shows a slight 
indentation. 

Little observable difference exists in the postcranial skeleton, except for the 
meristic counts. The caudal fin does not bear fringing fulcra which are present in 
the type-species. 



Davichthys lacostei (Arambourg) 

1954 Holcolepis lacostei Arambourg : 54, text-figs. 28, 29, pi. 7, figs, 1, 3, 4. 

Diagnosis (emended). Davichthys reaching 90 mm S.L. Skull roof smooth. 
Maxilla with convex oral margin, both supramaxillae large and rounded, premaxilla 
small. Preopercular sensory canal distant from the anterior margin of the pre- 
operculum. Dorsal fin nearer to the caudal peduncle than the occiput and composed 
of 14 rays. Anal slightly nearer to the caudal peduncle than the pelvic fins and 
composed of 9 rays. 



FOSSIL AND RECENT 35 

Formation and locality. Lower Cenomanian of Jebel Tselfat, Morocco. 

Remarks. This species has been described in some detail by Arambourg (1954). 
The chief differences from the other species concern the position and size of the median 
fins, the shape of the maxilla and the small size of the premaxilla. D. lacostei shares 
with D. gardneri a similarly shaped hyomandibular, an unornamented cranial roof 
and a similarly disposed preopercular sensory canal. 

Remarks on the genus Davichthys 

The close resemblance of Davichthys to other elopids has necessitated nothing 
more than the above comparative notes. Removal of the type-species from associa- 
tion with Lebonichthys gracilis is justified by the lack of any of the albulid specializa- 
tions characterizing Lebonichthys (p. 171). 

A new genus is made for these three species because they are intermediate in 
structure between the Upper Jurassic Anaethalion and the Eocene-Recent Elops. 
Davichthys appears to be the sole Cretaceous representative of the family Elopidae. 



Genus ANAETHALION White, 1938 
(Text-figs. 16-19) 

For generic synonymy see Gaudant (1968). For diagnosis of the genus see 
Nybelin (1967b). 

The genus Anaethalion as now understood is constituted by several Upper Jurassic 
(Kimmeridgian) species from Germany, France and Spain. The genus was first 
described by Munster (1842a) who used the preoccupied name Aethalion for his type- 
species, A. angustus. New species were added and described by Munster (op. cit. 
and 1842b), Winkler (1862) and Wagner (1863). Sauvage (1903) recognized two 
species (A. vidali and A. gigas) from Lerida, Spain. 

Nybelin (1967b), in a taxonomic revision of the genus as represented in the 
Kimmeridgian of Germany, recognized five (possibly six) species, viz. ; A . angustus 
(Munster), A. angustissimus (Munster), A. knorri (Blainville) , A. mayri Nybelin, 
A. sp. and A. (?) cf. subovatus (Munster). Lack of cranial material cast doubt on 
the inclusion of the last-mentioned species in the genus. Gaudant (1968) revised the 
species from France and came to the conclusion that in addition to A. knorri and 
A . angustus there were also two different species ; A . affmis Gaudant and A . cirinensis 
Gaudant. Both Nybelin (1967b) and Gaudant (1968) have described the anatomy 
of Anaethalion. 

The purpose of the present work is to evaluate the relationships of the genus, 
primarily upon the basis of the anatomy of Anaethalion vidali. No attempt is made 
to analyse the validity of the species recognized hitherto as the author has only had 
the chance to examine a limited amount of material. The material examined is as 
follows : Anaethalion angustus, B.M.N.H. 37926, 37912, 37927 ; A. knorri, B.M.N.H. 
36030, 379°3. 37 8 39. 37°42. P.1095 ; A. angustissimus, B.M.N.H. P-3575. 379 01 
(neotype) ; A. (?) cf. subovatus, B.M.N.H. P.3656, P.3723 ; A. sp., B.M.N.H. P.3728, 



36 ELOPIFORM FISHES 

37048 ; A.gigas, B.M.N.H. P.10927, P.10377, P. 10380 ; A. vidali, B.M.N.H. P.10375 
and P.10376. 

Anaethalion has been referred to the Leptolepididae by Woodward (1895), Berg 
(1940), Bertin & Arambourg (1958), Danil'chenko (1964) and Romer (1966). 
Nybelin (1967b) did not firmly suggest a familial position although he hinted strongly 
at a relationship with Elops. Similarly Gaudant (1968) could not reconcile Anaetha- 
lion with the Leptolepididae and erected a new family, the Anaethalionidae. 

I would agree with both Nybelin and Gaudant that Anaethalion should be 
excluded from the leptolepids. There are several features of dissimilarity from 
leptolepids which, at the same time, are points of similarity with Elops. The neuro- 
cranium is shallow ; there is a connection between the supraorbital and infraorbital 
sensory canals ; the supratemporal bone is large (unlike Upper Jurassic leptolepids) ; 
the infraorbital and preopercular sensory canals show few branches ; there is no 
basipterygoid process ; both the maxilla and dentary are shallow, the latter bearing 
only a weakly defined coronoid process posteriorly ; and the dentition upon the 
dermal jaws consists of broad bands of villiform teeth. 

Although Anaethalion is unlike leptolepids, it must also be said that definite 
evidence of elopid affinity is lacking for most species of Anaethalion. It is true 
that Anaethalion is most favourably compared with Elops and Davichthys, but a 
positive association with elopids could only be made on finding rostral ossicles, an 
associated leptocephalus larva or a pectoral splint. A pectoral splint is present in 
Anaethalion vidali and consequently this form may be assigned to the Elopidae with 
some degree of certainty (but see footnote p. 190). The other species may or may 
not be elopids. In caudal anatomy there is some variability from species to species 
within the genus Anaethalion. These caudal variations and their possible implica- 
tions in understanding the relationships of Anaethalion species are discussed below 

(P- 39)- 
The following remarks on the genus are based primarily on A. vidali, the only 

species definitely referable to the Elopidae. 

To describe the anatomy of A. vidali would essentially be a repetition of the 
description given for Elops, so alike are the two forms. The differences that exist 
are minor and many are bridged by Davichthys. 

The skull roof of Anaethalion differs from that of Elops only in the form of the 
dermethmoid, which bears ventro-laterally directed processes similar to those seen 
in young Tarpon and Megalops. Such horns are absent in Elops but may be seen 
in Davichthys gardneri and probably represent a primitive teleostean condition. 

Little is known concerning the neurocranium of any species of Anaethalion. 
What is known suggests that the neurocranium is primitive, closely resembling that 
of Elops. Within the otic region the pars jugularis is long with separate openings 
as in Elops, the subtemporal fossa is deep and there is a prootic-intercalar bridge in 
at least A. vidali and A. knorri. Posteriorly the first vertebral centrum is incor- 
porated with the neurocranium (also figured for A. knorri by Nybelin 1967b) and this 
centrum bears a large neural arch. 

The snout of A . vidali resembles that of Elops ; the mesethmoid must have been 
unossified and the lateral ethmoid is represented by a thin sliver of perichondral 



FOSSIL AND RECENT 37 

bone. Although the shape of the vomer could not be seen it is known to be a short 
element bearing many small villiform teeth. Other species of Anaethalion examined 
show a comparable development of the vomer and lateral ethmoid although most 
appear to show some endochondral ossification of the mesethmoid. 

In all species of Anaethalion the parasphenoid is relatively narrow throughout 
and is without a basipterygoid process. A. vidali differs from most species of the 
genus, and also from Elops, in the shape of the parasphenoid. In A . vidali the para- 
sphenoid is angled beneath the ascending wings. At this same level there are slight 
lateral swellings. Both of these features are seen in the parasphenoid of megalopid 
fishes and the shape of this element in A . vidali (Text-fig. 16, par) is most favourably 
compared with that of Tarpon (Text-fig. 23, par). Small villiform teeth are borne 
on the parasphenoid. 

Anaethalion vidali differs from other species of Anaethalion examined in having 
the otic region of the neurocranial roof convex and the autosphenotic spine well 
developed (Text-fig. 16, asp). 

The hyopalatine series is of the basic teleostean plan (Text-fig. 16). Unlike 
Davichthys the metapterygoid is relatively shallow and thus more closely resembles 
that seen in Elops. There is no evidence of an ectopterygoid process but the dorasl 
margin of the ectopterygoid is slightly thickened. The quadrate/mandibular 
articulation occurs beneath the posterior border of the eye, as in most other species 
of Anaethalion, and is in a similar position in Davichthys lacostei (Arambourg 1954 : 
fig. 28). In other species of Davichthys and Elops the articulation is situated more 
posteriorly. A. angustissimus shows a condition similar to Elops. The backward 
movement of the quadrate/mandibular articulation is a trend to be noted in the 
Elopidae. 

The upper jaw of A. vidali is similar to that of other elopids, showing a small 
premaxilla and small supramaxillae, the latter proportionately smaller than in 
Elops and Davichthys. The palatine process on the maxilla is relatively small in 
Anaethalion. Most species of Anaethalion (but not A . vidali) show a stout strengthen- 
ing ridge upon the posterior supramaxilla, a feature which is interpreted as being the 
primitive condition. 

The lower jaw is shallow as in other elopids and although other species of Anaetha- 
lion show a variation in the depth and shape of the mandible (see Nybelin 1967b) 
all show a weakly defined, posteriorly situated coronoid process, in contrast to lepto- 
lepids. 

The anterior ceratohyal of all examined species of Anaethalion is fenestrated, un- 
like that of Davichthys and Elops. A fenestrated ceratohyal is probably a primitive 
feature since it is found in a great many primitive teleosts. ~ 

The circumorbital series of A . vidali is not known but there are minor differences 
between other species of Anaethalion and Elops. The circumorbital series of several 
species of Anaethalion has been described and figured by Nybelin (1967b) and Gau- 
dant (1968). In those species of Anaethalion known in this respect the first infra- 
orbital is ovoid, with a triangular antorbital resting directly upon its antero-dorsal 
margin. There is no connection between the lachrymal and the supraorbital, and 
the triradiate division of the infraorbital sensory canal (i.e. the origin of the antorbital 



38 



ELOPIFORM FISHES 




FOSSIL AND RECENT 39 

and ethmoid commissure branches) lies within the lachrymal. In Elops the lachry- 
mal is elbowed and contacts the supraorbital dorsally, the splint-like antorbital 
lying along the anterior margin of the upright limb. The triradiate division of the 
infraorbital sensory canal occurs in the skin anterior to the lachrymal. Junction 
of the lachrymal and supraorbital bones, so excluding the antorbital from the orbital 
margin, is rare in lower teleosts, although it has been figured for Leptolepis dubia 
(Patterson 1967a : fig. 4) and is therefore of doubtful phyletic significance. The 
bone-enclosed branches of the infraorbital sensory canal of Anaethalion are similar 
to Davichthys but unlike Elops. 

The preoperculum of A. vidali and other species of Anaethalion consists essentially 
of a vertical limb, the horizontal portion being short. Elops, Davichthys dubius 
and D. lacostei show a more pronounced horizontal limb and the preoperculum is 
slightly expanded postero-ventrally. The preopercular sensory canal of all species 
of Anaethalion opens to the surface within the lower portion of the bone. In A. 
vidali, as in other elopids, the main canal opens directly to the surface via a series of 
pores but in other species of Anaethalion there are a few well-developed secondary 
branches enclosed within the bone. The latter condition is considered the more 
primitive. 

The postcranial skeleton of A . vidali is similar to that of Elops, with no significant 
differences in the vertebral column or the paired fins and girdles. The vertebral 
count of A. vidali (66) is substantially higher than that of most species of Anaethalion 
(range 42-57). 

In most species of Anaethalion the dorsal fin, although of the same shape and size 
as the dorsal of Elops, is situated slightly nearer to the caudal peduncle than the 
occiput. In Elops the dorsal is situated in the centre of the back. Among the 
species of Anaethalion, A. (?) cf. subovatus and A. vidali show an Elops-like dis- 
position of the dorsal fin. A. vidali exhibits three features of the fins normally 
associated with megalopids among Elopiformes : the origin of the pelvic is in advance 
of that of the dorsal, the anal fin is slightly elongated and the first anal pterygiophore 
is stout and long, nearly touching the vertebrae above. 

The caudal skeleton of A . vidali (Text-fig. 17) is very similar to that of Elops (or, 
for that matter, later leptolepids) the main difference being the development of a 
complete neural spine associated with the second preural centrum. However, a 
complete neural spine upon the second preural centrum is also seen in Davichthys 
gardneri among elopids. Unlike Elops, A . vidali shows a single fringing fulcrum, a 
structure which is also seen in Davichthys dubius. Anaethalion vidali differs from 
other elopids but resembles some other species of Anaethalion in the development of 
laminar bone on the anterior margins of the third and fourth preural neural spines. 

The caudal skeleton of the German species of Anaethalion shows several variations 
from that of A . vidali. Sketches of some of these species are given in Text-fig. 18 
(see also Nybelin 1971). The development of laminar bone varies ; it is well 
developed in A . knorri, A . angustus and also in the Spanish A . gigas ; A . angustissimus 
shows weakly developed laminar bone on the second preural neural spine and as in 
A. knorri and A. angustus there are well developed projections at the bases of the 
first few preural neural and haemal spines. 



4 o 



ELOPIFORM FISHES 



n pu1 



nsppu2 



e P 1 -3 un3 



pu5 




c.sc 



5 mm 
i 1 i 



Fig. 17. Anaethalion vidali (Sauvage). Caudal skeleton in left lateral view. Camera 
lucida drawing of B.M.N. H. P.10375, partially restored. Arrows indicate upper and 
lower principal fin-rays. 



There is also variation in the length of the second preural neural spine, long in 
A. angustus, ' three-quarter length ' in A. (?) cf. subovatus and ' half-length ' in A. 
angustissimus , A. knorri, A. gigas and A. sp. 

The three epurals of most species of Anaethalion form a graded series but those of 
A. angustissimus are subequal in length, as in many primitive euteleostean fish. 

It is also to be noted that A . gigas from the Kimmeridgian of Spain shows a large 
neural structure above the first preural and first ural centra. This appears to be a 
result of fusion between two neural arches which in other species remain separate. 
Such a fused structure may be a function of size. 

Regrettably, too few specimens were examined to allow an assessment of these 
variations. The neural arches and spines are known to vary in Megalops (Text-fig. 
33) and perhaps little attention is justified. The development of laminar bone in 
some species may be of significance since this feature is usually associated with the 
euteleostean caudal skeleton (Patterson 1970b). 



FOSSIL AND RECENT 



4i 



A 



B 




pu2 





D 




pu2 




Fig. 18. Anaethalion. Caudal skeletons of various species. Based on camera lucida 
drawings. A. Anaethalion (?) cf. subovatus from B.M.N.H. P. 3723 ; B. A. angustus 
from B.M.N.H. 37927 ; C. A. angustissimus from B.M.N.H. 37901 ; D. A. knorri from 
B.M.N.H. 37839 ; E. A. gigas from B.M.N.H. P. 10380. A to D have been reversed. 



42 ELOPIFORM FISHES 

The variations in the caudal skeleton of Anaethalion illustrated by Nybelin (1963, 
1967b, 1971) and here (Text-fig. 18) do not, however, encompass the type shown by 
Gaudant (1968 : fig. 5) for A. cirinensis, based on the holotype (Gaudant op. cit. 
pi. 3, fig. 2). The figure given by Gaudant is unusual for two reasons ; a complete 
neural arch and spine is shown on the first preural centrum and the third uroneural 
is elongate and overlaps the second ural centrum. Both these conditions are not 
met with in elopid (or late leptolepid) fishes, but are usually associated with 
osteoglossomorph fishes (cf. Hiodon and Lycoptera as figured by Greenwood 1970b). 

From the plate given by Gaudant the apparent discrepancy between his 
drawing and other species of Anaethalion may be reconciled. The neural structure 
borne by the first preural centrum appears to represent a neural arch only, there 
being no indication, even as an impression, of a distal prolongation. The uroneural 
structure is correctly figured as may be seen by comparing the plate and drawing and 
inasmuch as there are three visible uroneurals the labelling is accurate. However, 
the third uroneural figured by Gaudant probably represents the second of other 
Anaethalion species and elopoids. In length and proximal extension this uroneural 
is exactly comparable with the second uroneural of elopoids and other species of 
Anaethalion. Presumably the third uroneural is either missing from the specimen 
or possibly represented by the splint-like structure shown below the distal extremity 
of the ' third ' uroneural (Gaudant 1968 : pi. 3, fig. 2). The first two uroneurals 
(U^ and U 2 of Gaudant 1968 : fig. 5) represent the unfused condition of the first 
uroneural of elopoids which is known to be a compound element. Thus the first two 
uroneurals of A. cirinensis resemble those of leptolepids. 

The scales of Anaethalion (unknown in A. vidali) are moderately large, circular 
and marked by fine circuli which are absent from the central portion of the exposed 
field. Bone cells are present. Such features are similar to those of Elops. How- 
ever, the scales of Anaethalion do not show the well-marked anterior radii charac- 
teristic of all elopiforms. 

In summary, the genus Anaethalion is an assemblage of primitive teleostean fishes 
most favourably compared with the Elopidae. Anaethalion is more advanced than 
known leptolepids in showing a reduction of the ethmoid ossification and lacking 
a basipterygoid process. The suspensorium and dermal jaws are primitive as are 
many features of the neurocranium and give no indication as to the relationships of 
most species of Anaethalion. The caudal skeleton is seen to vary from species to 
species (Nybelin 1971) and it may be of significance that some species show laminar 
bone, a feature normally associated with euteleosteans. The structure of the caudal 
skeleton suggests that A. (?) cf. subovatus and A. sp. are closely related; A. 
knorri, A . angustus and A . angustissimus are probably interrelated species and may 
be linked to A. sp. and A. (?) cf. subovatus by A. vidali. Anaethalion cirinensis 
appears to be the most primitive species on the basis of caudal anatomy since it 
shows three long uroneurals, the first two representing a ' pre-fusion ' state compared 
with other species of Anaethalion. 

The elopid A. vidali resembles the Megalopidae in some respects. Thus the shape 
of the body, the position of the pelvic fin and the elongate first anal pterygiophore 
are similar to Tarpon. In the cranium the otic region of the cranial roof is convex, 



FOSSIL AND RECENT 



43 



w; 




Oh 

E 

s 

PQ 

a 

o 

-a 

<D 

en 

cS 

m 



T3 
4) 



C 
O 

-t-> 






D 
> 






s 

B 
S 






44 ELOPIFORM FISHES 

the autosphenotic spine is well developed and the dilatator fossa is relatively deep. 
These features, and the shape of the parasphenoid, are reminiscent of those seen in 
megalopids. Further research on A. vidali may justify the inclusion of this species 
in the Megalopidae. 

Family MEGALOPIDAE Jordan, 1923 

Diagnosis. Elopoid fishes in which the body is compressed. Cranium deep, 
mouth terminal or (more usually) superior. Cranial roof convex above the otic 
region. Parietals without evidence of middle pit-lines. Autosphenotic spine and 
epiotic process well developed. Post-temporal fossae extend forward to orbito- 
sphenoid and are confluent above the cranial cavity. Dilatator fossa with or without 
roof. Otophysic connection developed in at least some representatives ; intercalar 
enlarged, partially or completely surrounding the cranial diverticulum of the swim- 
bladder. One rostral ossicle. Quadrate/mandibular articulation beneath orbit. 
Mandible with prominent coronoid process. Pseudobranchiae present (Holstvoogd 
1965). Pelvic fins originating beneath or anterior to the dorsal fin origin. Anal fin 
slightly elongated (compared with Elopidae). First anal pterygiophore extending 
to the vertebral column. Urodermal absent. Scales with three to six basal radii, 
posterior field ornamented. Lateral line tubes branched. 



Genus TARPON Jordan & Evermann, 1896 

For full generic synonymy see Hildebrand (1963). 

Diagnosis (emended). Megalopid fishes reaching 1800 mm in length. Neuro- 
cranial roof markedly convex above the cranial vault. Parietals considerably longer 
than broad in adult. Dilatator fossa broad, shallow and without a roof. Intercalar 
not entering the lateral wall of the periotic bulla. Jugular canal opening below the 
level of the subtemporal fossa. Maximum depth of the neurocranium at the occiput. 
Maxilla extending behind the eye. Scales with three to four anterior (basal) radii. 

Type and only species. Megalops atlanticus Cuvier & Valenciennes, 1846. 

Tarpon atlanticus (Cuvier & Valenciennes) 
(Text-figs. 20-29) 
For synonymy see Hildebrand (1963). 

Diagnosis (emended). Tarpon in which the origin of the pelvic fin is in advance 
of the dorsal fin. Maximum depth of body less than the length of the head. Dorsal 
fin with 13-15 rays, anal 22-25. Vertebral column with 53-57 vertebrae of which 
approximately 20 are caudal. Lateral line with 41-48 scales. Fringing fulcra one 
to four in number. 

Habitat. Atlantic coastal waters of North and South America (from Cape Cod 
to Brazil), sometimes entering fresh water. Also recorded from tropical West Africa. 



FOSSIL AND RECENT 45 

There is a great deal of literature on both locality data and growth stages of T. 
atlanticus, a useful synopsis of which is found in Hildebrand (1963). 

Remarks. The skeleton of Tarpon atlanticus has never been satisfactorily de- 
scribed. This, and the fact that in several respects it is more primitive than the 
frequently cited Megalops cyprinoides has led to the inclusion here of a more complete 
account. Omitted details are those of the caudal fin anatomy, which have been dealt 
with by Hollister (1936 : 263-268), and the otic region of the skull (in so far as this 
has been affected by the anterior swimbladder diverticula) which has been described 
by Greenwood (1970a). 

Description. In a young individual (c. 87 mm S.L.) the cranium is one and a 
half times as long as deep. Throughout life the depth increases with respect to 
length. Other morphometric changes that occur during growth are dealt with in 
the description. Unless otherwise stated, remarks apply to adult individuals. 

Neurocranium. The skull roof is flat anteriorly but becomes strongly convex 
posteriorly. In lateral view the dorsal margin of the skull roof is perfectly straight, 
or in some large specimens may even exhibit a slight concavity (cf. Gregory 1933 : 
fig. 31). The greatest width of the neurocranium is equal to 60 per cent of its 
length and occurs at the sphenotic level. During growth the skull roof becomes 
relatively wider, due primarily to the development of strong sphenotic spines, and the 
distance from the occiput to the level of the sphenotic spines increases relative to 
the total neurocranial length. This differential growth is quite considerable (in a 
fish of c. 87 mm S.L. the distance from the occiput to the level of the sphenotic 
spines is 26-5 per cent of the total neurocranial length, while in a fish of 650 mm this 
distance is 37 per cent) and may reflect the development of the enormous post- 
temporal fossae. 

The anterior portion of the skull roof is formed, as in all megalopids, by a complex 
median dermethmoid. The horizontal part of this bone is constricted one-third of 
the way back. To the anterior margin of the bone is attached a strong ligament 
from the head of the maxilla. The ethmoid commissure runs transversely through 
the tubular rostral, near the anterior border. Laterally, a large pore receives the 
canal from the lateral rostral. Immediately medial to this, another large pore 
opens posteriorly and near the mid-line a much smaller pore opens anteriorly. 
Posteriorly, the dermethmoid overlaps the anterior end of the frontals. A postero- 
ventrally directed wing arises from beneath the lateral border of the dermethmoid 
and passes down over the cartilage surrounding the olfactory organ. Posteriorly 
this wing, in older specimens, is in sutural contact with the lateral ethmoid. 

It appears that the lateral processes are ossifications embryonically distinct from 
the main body of the dermethmoid. Patterson (1970b) states that similar ossifica- 
tions are present in Leptolepis coryphaenoides and Megalops and that they are similar 
to the ' proethmoids ' of the stomiatoid Polymetme. For a fuller account of the 
homology of these lateral processes see Patterson (1970b : 261-262). 

The frontal is the largest single component of the skull roof. Anteriorly it passes 
beneath the dermethmoid almost to the anterior end of the latter (Nybelin 1967a : 
fig. 2C figures the anterior termination of the frontal as lying completely behind the 
dermethmoid), while posteriorly the frontal overlies part of both the pt erotic and 



4 6 



ELOPIFORM FISHES 

e com 



mes 



so.sc 



asp 




pa — 



otsc 



pto 



stt.com 



Fig. 20. Tarpon atlanticus (Cuvier & Valenciennes). Neurocranium in dorsal view. 
Nasal, dermosphenotic and supratemporal are shown on the right side only. This and 
Text-figs. 21-27 are eacn based on several adult specimens. Hatched areas represent 
cartilage. 



FOSSIL AND RECENT 47 

parietal. The surface of the frontal is marked by a very shallow ridge which runs 
the length of the bone and extends posteriorly on to the parietal. 

Anteriorly the supraorbital canal is contained within a horizontal plate-like 
nasal. In smaller individuals the nasal is tubular, as in Elops. Above the orbit 
the frontal has a narrow rebate for the narrow supraorbital. In larger specimens 
the supraorbital is tightly bound to the frontal. 

The parietal is three times as long as wide and is overlapped by the frontal an- 
teriorly. During growth the shape of the parietal changes from broad to narrow. 
The pterotic is triangular and its true extent may not be seen in dorsal view, due to 
the convexity of the skull roof. 

Beneath the dermethmoid the ethmoid cartilage extends back to the orbit. 
Anteriorly, where the cartilage projects beyond the dermethmoid, it becomes ossified 
as it expands over the dorsal part of the vomer. This ossified mesethmoid is securely 
attached to the vomer, so that its individuality may only be recognized in young, 
alizarin-stained specimens. The dorsal surface of the mesethmoid bears a median 
ridge on either side of which is an ovoid depression. The depression receives 
the head (cranial condyle) of the maxilla, while the premaxillae meet one another 
immediately above the median ridge. 

The medial wall of the nasal capsule is mainly cartilaginous. The lateral ethmoid 
is ossified in the lateral part of the planum antorbitale. From its sutural connection 
with the underside of the frontal the lateral ethmoid curves anteriorly and ventrally 
to contact the dermethmoid wing. The ventral surface of this contact bears a large 
cartilaginous meniscus which receives the dorsal surface of the palatine head. 

The vomer underlies the mesethmoid and the anterior end of the parasphenoid. 
In ventral view it is diamond-shaped with the posterior angle produced in an 
elongated point which interdigitates with the parasphenoid. The dorsal surface 
of the vomer contributes to the facet receiving the maxillary head. The vomer 
bears an ovoid patch of numerous fine, backwardly pointed teeth. The confluence 
of paired vomerine tooth plates is more complete in adult Tarpon than in Elops, 

From the vomer the parasphenoid narrows beneath the orbit before expanding 
at the level of the ascending wings. Posteriorly the parasphenoid tapers and 
divides into two prongs which end below the posterior part of the basioccipital. 
The gap between these two prongs represents the anterior part of a groove beneath 
the basioccipital and the first vertebral centrum. The dorsal aorta runs in this groove 
and divides into the lateral dorsal aortae at its anterior end. Such a groove is clearly 
necessary in this form since the swimbladder diverticulum is pushed tightly against 
the skull base beneath the basioccipital. Beneath the ascending process the 
parasphenoid is produced laterally in a small prominence at the apex of which the 
internal carotid artery enters the myodome. Immediately anterior to this promi- 
nence a shallow groove curves antero-ventrally and marks the cranial origin of the 
adductor arcus palatini musculature. A very narrow band of small granular teeth 
is found on the parasphenoid beneath the orbit. In all megalopids the parasphenoid 
tooth patch is less extensive than that of the Elopidae. 

The pterotic forms the posterior part of the hyomandibular facet and the dorsal 
region of the subtemporal fossa. The posterior face of the pterotic forms the roof 



4 8 



ELOPIFORM FISHES 

20mm soc 



exo 




pap 
Fig. 21. Tarpon atlanticus (Cuvier & Valenciennes). Neurocranium in posterior view. 



and lateral border of the post-temporal fossa. The exoccipital is large, forming the 
ventral border to the post-temporal fossa opening, and the dorsal and lateral margins 
of the foramen magnum. Along the medial border of the post-temporal fossa the 
exoccipital forms a ridge which is continued on the epiotic. Medial to this ridge 
both bones exhibit concave posterior faces. Thus the surface topography of these 
bones results in there being a median depression flanked on either side by the post- 
temporal fossae. Beneath the foramen magnum the exoccipital bears a facet which 
articulates with the first vertebral centrum. On the lateral face of the exoccipital 
and basioccipital there is a prominent postero-ventrally inclined ridge which provides 
a point of attachment for the anterior diverticulum of the swimbladder (Greenwood 
1970a has described and figured the extension of the swimbladder on the otic region 
of the skull). Baudelot's ligament inserts on the posterior face of the basioccipital 
part of the ridge. The exoccipital also contributes to the subtemporal fossa. 
Beneath the subtemporal fossa the exoccipital exhibits a marked depression. Into 
the upper part of this depression (which is covered in lateral view by the intercalar, 
not described here, see Greenwood 1970a) opens the vagus foramen, immediately 
preceded by the much smaller glossopharygeal foramen. The jugular vein runs 
along the roof of this depression. 

Beneath the exoccipital the basioccipital forms the inner wall of the periotic bulla. 
The ontogenetic changes that affect the basioccipital are discussed by Greenwood 
(1970a). 



FOSSIL AND RECENT 



49 



VO 



ptfc 



pro. ic 



ic 




fa hm 



exo 



fuv 



10mm 



Fig. 22. Tarpon atlanticus (Cuvier & Valenciennes). Neurocranium in ventral 
view. Hatched areas represent cartilage. 



50 ELOPIFORM FISHES 

The posterior face of the basioccipital and the exoccipital provide a tripartite facet 
for the first vertebral centrum, which is a functional part of the neurocranium. 
The centrum bears autogenous parapophyses and a neural arch. 

The prootic is, as usual, of complex shape. The lateral face contributes to the 
hyomandibular facet, the subtemporal fossa and the medial wall of the periotic 
chamber. A stout ridge runs horizontally across the lateral face of the prootic 
beneath the subtemporal fossa. The jugular canal opens near the posterior border 
of the prootic, immediately beneath the horizontal ridge. The hyomandibular 
foramen is large and points dorsally above the pars jugularis. In some (particularly 
large) specimens an additional foramen (Text-fig. 23) is situated dorsal to the 
hyomandibular foramen (Greenwood 1970a : pi. 3, fig. 1). The significance of this 
foramen is not clear ; it may provide a separate opening for the efferent hyoidean 
artery. Another foramen of inconsistent occurrence may be seen immediately 
posterior to the hyomandibular foramen. Greenwood (1970a) identifies this as ' the 
foramen for the head vein ' but this is unlikely since in the specimen figured (Green- 
wood 1970a : pi. 3, fig. 1) there is a foramen for the head vein in the usual position 
and also the foramen in question has no connection, direct or indirect, with the pars 
jugularis but opens into the post-temporal fossa. Dissection of a specimen showing 
a foramen in a similar position failed to reveal any function. A palatine branch of 
the trigeminal nerve leaves the prootic by a slit-like foramen beneath the level of 
the pars jugularis. The orbital artery pierces the prootic near its suture with the 
parasphenoid. 

The medial wall of the pars jugularis is pierced by two large foramina. The pos- 
terior and more ventral of these is the facial foramen while the anterior is the tri- 
geminal foramen (the profundus ciliaris also passes through this foramen). The 
anterior face of the prootic contacts the autosphenotic and pterosphenoid dorsally 
and the parasphenoid and basisphenoid ventrally. The anterior face slopes postero- 
ventrally from its dorsal contact with the pterosphenoid. Ventrally the prootic 
turns horizontally meeting its partner in the mid-line to form the prootic bridge. 
The abducens nerve pierces the bridge. 

The pterosphenoid is inclined almost horizontally. The trochlearis leaves the 
endocranial cavity through a small foramen in the prootic. It passes into the 
pterosphenoid to run forward within the bone, leaving the pterosphenoid at its 
anterior limit. The anterior margin of the orbitosphenoid is notched by the exit 
of the olfactory tract. The orbitosphenoid forms the anterior wall of the post- 
temporal fossae. 

The median basisphenoid has a stout pedicel which in older specimens becomes 
firmly attached to the parasphenoid. From the posterior border two postero- 
laterally inclined prongs enter the posterior myodome cavity. The median basi- 
sphenoid pedicel bifurcates dorsally to form two short wings which contact with the 
prootics. 

The post-temporal fossae are enormous. The posterior opening of each fossa 
is deeper than wide (cf. Megalops). The fossae extend right forward to the 
anterior end of the orbitosphenoid and meet in the mid-line above the endocranial 
cavity, from which they are separated by cartilage. The increased depth of the 



FOSSIL AND RECENT 



5i 




a! 



O, 



a! 



a) 
u 



o 



c 
a 

CD 

'o 

c 

> 



> 



s 

s 

o 







52 ELOPIFORM FISHES 

skull at the occiput is due to the dorsal development of these fossae. Epaxial trunk 
musculature occupies the fossae and is inserted on the walls and on a stout 
ligament which inserts on the medial surface of the autosphenotic. 

Hyopalatine bones. This series is almost as deep as long in adult specimens, but 
in younger specimens the series is considerably longer than deep. Correlated with 
this deepening during growth there is a slight forward shift in the quadrate/mandibu- 
lar articulation. In the adult it lies beneath the posterior half of the orbit. 

The hyomandibular is vertical. Its head has anterior and posterior expansions. 
The external surface of the bone is marked by a prominent ridge which ends dorsally 
in a small dorso-lateral projection. Immediately behind this ridge there is a groove. 
The hyomandibular nerve penetrates the medial face immediately beneath the head 
and divides into an opercular branch and a combined mandibular and hyoidean 
branch. The former branch leaves the hyomandibular on a level with the opercular 
process. The latter opens into the posterior groove immediately beneath the oper- 
cular process. The anterior border of the hyomandibular is produced into a narrow, 
very thin wing of bone. 

The long symplectic is inclined forwards at about 45 ° and fits into a deep groove 
upon the quadrate. The dorsal end of the symplectic overlaps the preoperculum. 
The quadrate is the usual fan-shaped bone. Ventrally there is an enlarged, trans- 
verse condylar surface behind which, in larger specimens, there is a shallow notch. 
Such a notch may limit the downward movement of the mandible. The anterior 
border of the quadrate fits into a groove on the ectopterygoid. 

The metapterygoid is separated from the dorsal margin of the quadrate by a thin 
slip of cartilage. Upon the lateral face the metapterygoid is produced postero- 
dorsally as a thin wing which overlaps the ventral part of the hyomandibular. 
Between this overlap of the hyomandibular and metapterygoid there is a consider- 
able space, the hyomandibular-metapterygoid cup, into which insert the deeper 
divisions of the levator arcus palatini muscles. 

The endopterygoid is overlapped posteriorly by the metapterygoid. Anteriorly 
the endopterygoid is ovoid in outline and fits between the posterior region of the 
auto- and dermopalatine. The ectopterygoid is a boomerang-shaped element, the 
two limbs being equally developed. The posterior and dorsal margins are grooved 
to receive the quadrate and endopterygoid. 

The dermopalatine is a flat plate of bone partially overlapping both the endo- 
and ectopterygoid. The medial edge of the dermopalatine lies against the border 
of the vomerine tooth patch. An autopalatine may ossify in large individuals. An- 
teriorly the palatine cartilage is expanded into a large knob, the dorso-medial face 
of which contacts the cartilage underlying the suture between the dermethmoid and 
lateral ethmoid. The antero-lateral aspect of the knob articulates with the palatine 
head of the maxilla. Between these two elements lies a small biconcave meniscus 
of cartilage. 

Much of the buccal face of the hyopalatine series bears teeth. Those on the pala- 
tine are small, pointed and directed posteriorly. Upon the endopterygoid, ecto- 
pterygoid (and, in older specimens, the metapterygoid) the teeth are slightly smaller 
and more granular in form. In young individuals the teeth are relatively larger, 



FOSSIL AND RECENT 



53 



E 




"3 . 

tn .5 

i S 



c/> o 

c & 

■r-i *0 

I I 

<n ™ 

■r W) 

> (D 

3 +J 

n <n 

u s 

■S <* 

« 8 

H p 






54 ELOPIFORM FISHES 

pointed, but much more sparsely distributed. Those on the ectopterygoid are 
confined to the extreme lateral border. 

Dermal upper jaw. The upper jaw ends behind the eye. In young individuals 
the jaw is relatively narrow with a slightly curved oral margin but in adult specimens 
the jaw is deeper and the oral margin is markedly rounded. 

The premaxilla occupies 20 per cent of the total jaw length. Anteriorly the 
premaxilla is deep, meeting its partner in the mid-line. The dorsal border is pro- 
duced posteriorly as a rudimentary articular process which overlaps the maxilla. 
The oral border of the premaxilla bears many small, pointed teeth. 

The maxilla is thin posteriorly but thickens anteriorly as it curves medially to 
articulate with the ethmoid region by a slightly expanded head. A small palatine 
process arises from the lateral face of the maxilla, beneath which there is a groove 
receiving the posterior (dorsal) border of the premaxilla. Two supramaxillae over- 
lap the dorsal margin of the maxilla. The smaller anterior supramaxilla is ovoid 
while the larger posterior element has a spine passing dorsal to the anterior supra- 
maxilla. A broad band of teeth is borne by the maxilla. The greater width of this 
tooth band is on the medial side of the bone. 

Mandible. Although the total length of the mandible equals that of the upper 
jaw, the mandible protrudes in front of the premaxilla and the mouth is upwardly 
directed. The profile of the dentary symphysis continues that of the skull roof in 
adult specimens, but in young individuals this is not so and the dentary symphysis 
is relatively shallow. 

The dentary forms two-thirds of the mandible. The ventral margin is slightly 
inflected. Anteriorly the dentary curves markedly inwards to the symphysis. For 
one third of the mandibular length the oral margin of the dentary is parallel to the 
ventral border, but behind this level the profile rises steeply to form much of the 
elongate coronoid process. In young individuals the dentary is relatively shallower 
and the oral margin is gently curved. Teeth are borne in a broad band along the 
oral margin of the dentary. The tooth band increases in width posteriorly where it 
spreads over the medial surface of the bone. Teeth borne by the dentary are similar 
to those of the upper jaw and dermopalatine. On the medial surface of the dentary 
there is a deep, anteriorly directed, crescent-shaped fossa which receives the anterior 
end of the A w muscle. 

The articular forms the posterior part of the coronoid process and the outer portion 
of the articulatory cup. The medial surface of the articular has a prominent ridge 
which is continuous with a ridge upon the dentary. The endosteal articular forms 
the inner portion of the articulatory cup and sends a short process forward on the 
top of the articular ridge. A small sesamoid articular occurs anterior to the endo- 
steal articular. 

Circumorbital series. This series is composed of seven bones, one of which, the 
supraorbital, is an anamestic component. The antorbital is a narrow strut, inclined 
antero-ventrally from its loose connection with the supraorbital. Part of the ventral 
antorbital margin lies against the dorsal border of the lachrymal (the first infra- 
orbital) . The latter element is rounded anteriorly. Posteriorly it narrows where it 
is partially overlapped by the second infraorbital. On the antero-dorsal part of the 



FOSSIL AND RECENT 

in 



55 



Tp ce 




grah 




dhh 



brr 

Fig. 25. Tarpon atlanticus (Cuvier & Valenciennes). Hyoid bar and branchiostegals 
(except uppermost ray) of right side in A, medial view and B, lateral view. 



first infraorbital there is a prominent knob articulating with the maxilla. Infra- 
orbitals 3, 4 and 5 form a broad, thin bony plate. The dermosphenotic is a narrow 
bone which tapers as it curves anteriorly to almost touch the supraorbital. It is 
considerably stouter than any other infraorbital. 

Hyoid arch, gill arches and gular plate. The ceratohyal, as usual, comprises 
anterior and posterior elements which are united by cartilage. A short, stout inter- 
hyal is connected to the posterior ceratohyal by cartilage. The upper and lower 
hypohyals are synchondrally united to one another and to the anterior ceratohyal. 
The medial surface of the lower hypohyal bears a small ovoid facet to which a liga- 
ment from the urohyal is attached. A medial prominence on the upper hypohyal 
lies against the basihyal. The afferent hyoidean artery pierces the medial surface 
of the lower hypohyal. From here the artery passes into the upper hypohyal, 
emerges from its dorsal surface, and runs posteriorly in a groove on the lateral face 



56 ELOPIFORM FISHES 

of the anterior ceratohyal. A foramen on the medial surface of the upper hypohyal 
connects with the arterial canal : the significance of this foramen is not clear. 

A tooth plate is associated with the dorsal margin of the anterior ceratohyal and 
part of the upper hypohyal. Additional tooth plates are scattered in the skin 
covering the medial surface of the anterior ceratohyal. 

There are 21-24 branchiostegal rays. The anterior 13 are thin and pointed 
(almost filiform) and attached to the ventral margin of the anterior ceratohyal 
which is scalloped for their reception. As one passes back along the series the rays 
become spathiform (terminology after McAllister 1968) and are attached to the 
lateral surface of the posterior ceratohyal. The upper branchiostegal extends further 
forward than those preceding it and is more closely associated with the interoper- 
culum. 

The gill arch structure is primitive for a teleostean fish. For instance, the basi- 
branchial complex is formed by independent dermal and endochondral elements 
and the tooth plates associated with the cerato- and epibranchials are very small and 
numerous. 

The basibranchial complex is formed by an ossified basihyal succeeded by three 
ossified basibranchials. The first basibranchial is short and deep and set slightly 
below the level of the elongate second and third elements. The fourth basibranchial 
is represented by cartilage. Tooth plates associated with the basihyal and basi- 
branchial are an elongated basihyal plate succeeded by a long tooth plate overlying 
the first three basibranchials. The tooth plate overlying the fourth basibranchial 
varies in shape and size from specimen to specimen : in young adults there are two 
ovoid tooth patches set one behind the other, but in old specimens only a single plate 
is recognizable. Thus, throughout life there is some degree of assimilation (Nelson 
1969a : 484). 

Hypobranchials 1-3 are attached to the posterior edge of their respective basi- 
branchial supports. There are five ceratobranchials and four epibranchials, the 
fourth epibranchial with a large expansion posteriorly. There is no fifth epibranchial 
as figured for Megalops by Holstvoogd (1965 : fig. 3b). Of the four infrapharyngo- 
branchials the anterior three are ossified and the fourth is represented in cartilage. 
The first infrapharyngobranchial articulates with the parasphenoid. Supra- 
pharyngobranchial 1 is ossified and attached to the dorsal limb of epibranchial 1 by 
cartilage. 

Tooth plates of the first arch are associated with hypo-, cerato-, epi- and infra- 
pharyngobranchials. Gill-rakers are present, associated with all the above elements 
except the first and fourth infrapharyngobranchials. On the first arch there is a 
total of 62 gill-rakers (16 on the hypobranchial, 24! upon the ceratobranchial, 20^ 
on the epibranchial and 1 lying free between the epibranchial and infrapharyngo- 
branchial). Immediately behind the gill-rakers and resting upon a ledge near their 
base there is a single row of rectangular tooth plates. On the ceratobranchial and 
much of the epibranchial there is a one-to-one correspondence between tooth plate 
and gill-raker. At the top of the epibranchial the tooth plate row breaks into 
smaller irregular plates which continue in the dermis overlying the infrapharyngo- 
branchial. Over the hypobranchial the tooth plate row has consolidated into larger 



FOSSIL AND RECENT 



57 



stt.com 



pmx 




smx 1-2 



qu brr 



sop 



70mm 



Fig. 26. Tarpon atlanticus (Cuvier & Valenciennes). Cranium in left lateral view. 



units. The posterior face of the hypo-, cerato- and epibranchial is covered by two 
rows of tooth plates. The more dorsal of these is formed by very small plates, 
irregular in shape and not always set in a linear fashion (indeed this ' row ' is best 
described as a band of plates). The ventral row is more easily definable with the 
tooth plates larger and regular in shape and set in a definite row. At the dorsal 
end of the epibranchial the upper row of tooth plates becomes a broad band which 
runs in continuity with the ventral row. 

The tooth plates associated with the second and third arches are set in similar 
fashion to those of the first. The tooth plates associated with the second infra- 
pharyngobranchial are larger than those on the first. The third infrapharyngo- 
branchial bears a well-developed tooth plate which lies in series with that associated 
with infrapharyngobranchial 4. 

The tooth plates upon the fourth arch are similarly distributed to those on the first 
three, but the most posterior row is formed by plates resembling reduced gill-rakers 
and these inter digitate with the reduced gill-rakers upon the fifth ceratobranchial. 
The three tooth-plate rows may still be recognized on the fifth ceratobranchials but 
the plates of the posterior row have become much enlarged and together form the 
lower pharyngeals. 

The urohyal extends posteriorly to the level of the fourth basibranchial. An- 
teriorly it is attached to the lower hypohyal of either side by a short ligament. The 
gular plate is large, extending for at least half the length of the mandible. 



58 ELOPIFORM FISHES 

Opercular series. The greatest width of the operculum occurs at the level of the 
articulation with the hyomandibular. Above the articulation the anterior margin 
of the operculum is concave. The posterior border of the bone is smoothly rounded. 
Postero-ventrally the ventral margin is obliquely inclined. The suboperculum is 
essentially a parallel-sided bone with rounded anterior and posterior margins. The 
interoperculum is large and rhomboidal in shape. It is attached to the articular 
by a stout ligament. 

The preoperculum has a smoothly concave anterior border. Mid-way along its 
length there is a small prominence anteriorly for the insertion of part of the adductor 
mandibulae muscle. The preoperculum is expanded postero-ventrally. 

Cephalic sensory canal system. The supraorbital canal penetrates the frontal 
above the autosphenotic spine and curves anteriorly to run within the shallow ridge. 
A large backwardly directed branch of the canal runs within the posterior part of the 
ridge to open on to the parietal. Within the frontal the main secondary branches 
of the canal are as illustrated (Text-fig. 20) but in younger individuals the branches 
are far more numerous. 

A large otic sensory canal runs along the lateral border of the pterotic 
opening anteriorly above the autosphenotic spine and posteriorly above the posterior 
limit of the hyomandibular facet. 

The infraorbital sensory canal runs near to the orbital margin of all infraorbitals 
except the first where it passes diagonally across the bone to open at the base of the 
antorbital. Short branches are given off posteriorly from the main canal. Within 
the skin these canals extend for a considerable distance and exhibit prolific branching. 
The infraorbital canal bifurcates as it leaves the anterior (dorsal) end of infraorbital I, 
sending a thin antorbital branch (adnasal branch of some authors, e.g. Nybelin 
1967a) backwards, and a larger canal forwards which runs through a rostral ossicle 
before uniting with the ethmoid commissure (this region of the canal system has been 
excellently described and figured by Nybelin 1967a). The infraorbital canal meets 
the supraorbital and otic canals within the skin above the dermosphenotic. 

The preopercular sensory canal runs very close to the anterior border and opens 
on to the surface of the bone through four to seven pores. 

The mandibular sensory canal pierces the posterior surface of the articular and 
runs within this bone and the dentary. The canal opens to the surface by pores 
which become more numerous anteriorly. 

Pectoral girdle and fin. The supratemporal is very similar to that of Elops 
hawaiensis (p. 22). The pattern of the contained lateral line and supratemporal 
commissure is also similar to Elops, with the difference that the commissure in Tarpon 
has very few pores. 

The post-temporal is a large concavo-convex bone disposed in the vertical plane. 
The anterior margin shows three projections, separated by two shallow indentations. 
The medial (dorsal) projection overlies the dorso-lateral face of the epiotic process 
to which it is attached by a ligament. The middle process is wide and truncated 
anteriorly. To this limb is attached a stout ligament which runs into the post- 
temporal fossa. The spine-like lateral (central) process is inclined antero-ventrally 
and connects with the basal part of the intercalar by a stout ligament. 



FOSSIL AND RECENT 



59 




pel 2 



cor 



Fig. 27. Tarpon atlanticus (Cuvier & Valenciennes). 

side in medial view. 



Pectoral girdle of right 



The lateral line canal descends from the underside of the supratemporal and runs 
within a bony tube on the outer surface of the post-temporal before leaving the bone 
to pass into the supracleithrum. Posteriorly the post-temporal has a small promi- 
nence which fits into a depression at the dorsal end of the supracleithrum. 

The blade-shaped supracleithrum is pierced dorsally by the lateral line canal which 
runs obliquely across its lateral face. Numerous ventral branches are given off 
from the supracleithral canal. 



60 ELOPIFORM FISHES 

The cleithrum is large and curves antero-ventrally. The anterior tip meets its 
fellow in the mid-line. The anterior margin of the bone is thickened and dorsally 
this thickening forms a short spine to which Baudelot's ligament is attached. A 
horizontal ledge projects medially from the ventral part of the cleithrum. 

The endoskeletal girdle consists of the scapula, coracoid and mesocoracoid. The 
scapula is of complex shape. Laterally it spreads over the medial face of the cleith- 
rum. Medially, the horizontal portion of the scapula contacts the coracoid and meso- 
coracoid through synchondral unions. The posterior margin of the scapula bears a 
saddle-shaped depression receiving the first pectoral fin-ray. The scapular foramen 
is incomplete. 

The coracoid is elongate and lies in the vertical plane. Posteriorly the lateral 
face unites with the scapula through a simple synchondral join, and with the meso- 
coracoid through an interdigitating synchondral join. The coracoid also exhibits 
a syndesmotic union with the horizontal cleithral ledge. The posterior margin of 
the coracoid is capped by cartilage. The medial face of the coracoid lies against 
that of its partner while the anterior margin is bound by loose connective tissue to 
the cleithrum. A large foramen pierces the coracoid below its union with the 
scapula and mesocoracoid. The latter bone spreads over the inner surface of the 
cleithrum. A narrow mesocoracoid arch passes ventro-medially to contact the 
scapula and coracoid. 

There are four ossified radials. The outermost is the shortest and the only member 
of the series attached to the scapula. The other members of the radial series are 
attached to the coracoid, the innermost being the longest. There are two post- 
cleithra. The dorsal one is rounded and scale-like. Anteriorly this element bears 
a small prominence which lies medial to the cleithrum. The ventral postcleithrum 
is an elongated curved roof. The upper end is somewhat flattened and lies against 
the medial edge of the dorsal postcleithrum. 

There are 14 pectoral fin-rays, the outermost unbranched. All rays are segmented. 
At the base of the upper half of the first fin-ray there is a triangular splint bone. 

Vertebral column. There are approximately 54 vertebrae (range 53-57) of which 
about 20 are caudal. All centra are amphicoelous and pierced by the notochord. 

Most of the centra are isodiametric but those anteriorly are deeper than long. 
Each centrum is marked laterally by numerous grooves separated by fine ridges, 
some of which anastomose. The neural arches are autogenous. Each centrum 
bears pits dorsally for their reception. The neural arches associated with the first 
29 centra are expanded distally. Neural spines borne by these arches are slender, 
straight and remain separate from their fellows of the opposite side. The vertebral 
centrum which is incorporated into the neurocranium bears a neural arch somewhat 
broader than those immediately behind, but the neural spine is considerably shorter. 
The neural spines of the posterior abdominal and caudal region are stout and fused 
to their partners of the opposite side. The neural arches which support caudal 
fin-rays are modified for this function and are not considered here (see Hollister 

1936). 

All of the caudal centra have stout haemal arches and spines. The posterior 
haemal spines are modified in support of the caudal fin-rays. Autogenous 



FOSSIL AND RECENT 



61 



un1-2 



ep1-3 



nsp pu3 



pu 3+4 




Fig. 28. Tarpon atlanticus (Cuvier & Valenciennes). Caudal skeleton in left 
lateral view. Based on a young individual 87 mm S.L. 



parapophyses are found throughout the abdominal region. Those upon the last 
eight abdominal centra become progressively longer. Pleural ribs are attached to 
the ventro-posterior aspect of all parapophyses, including those associated with 
the neurocranial centrum. 

Epineurals are attached to the bases of the neural arches throughout the abdominal 
region. In the caudal region the dorsal series of intermuscular bones have forked 
bases which are attached to the bases of the neural arches. Epipleural (ventral) 
intermuscular bones attach to the dorsal surface of the parapophyses. In the 
caudal region ventral intermusculars are attached to the bases of the haemal arches 
by tough connective tissue. The ventral intermuscular bones of the anterior 12 
vertebrae are attached to the lateral face of the centrum and there are small pro- 
jections upon the centra to receive the intermuscular heads. Thus in the extreme 
anterior region the intermusculars are epicentral in position. An intermuscular 
bone which is attached to the exoccipital appears to lie in series with these epicentrally 
placed intermusculars. In the caudal region fine intermuscular bones are attached 
to the centre of each centrum by a narrow strand of connective tissue and are there- 
fore to be regarded as epicentrals. Thus there are three series of intermusculars 
in the caudal region ; dorsal, epicentral and ventral intermusculars. 

Between the neural spines of the first 22 vertebrae there is a series of 23 supra- 
neurals. All but the first are slender and tubular. The first is solid, very broad, 



62 ELOPIFORM FISHES 

particularly in its dorsal extent, and situated anterior to the reduced neural spine 
associated with the neurocranial centrum. As in Elops (p. 25) there is an overlapping 
of the supraneural and pterygiophore series. 

Median fins. The dorsal fin is situated in the centre of the back and is composed 
of 15 (13-15) fin-rays, of which the first two are unsegmented. The sixth ray is the 
first branched. Behind the sixth the rays decrease in length except the last, the 
posterior half of which is produced into a long filamentous structure (see p. 201). 
The lepidotrichs are supported by 12 pterygiophores, each composed of proximal, 
middle and distal radials. The first pterygiophore is considerably stouter than 
those succeeding it, and proximally it extends between the tips of the neural arches 
of the twenty-first and twenty-second vertebrae. 

The origin of the anal fin lies nearer to the pelvic than to the caudal peduncle. 
The fin is composed of approximately 24 rays (range 22-25), 0I which the first four are 
unsegmented. The eighth is both the longest and the first branched ray. The 
posterior rays increase in length. The posterior half of the last ray may be produced 
as a short filament (in alcohol-preserved specimens the tips of the fin-rays are often 
broken). The fin-rays are supported upon approximately 21 pterygiophores. The 
first, which supports four rays, extends antero-dorsally to end immediately beneath 
the thirty-first centrum. 

The caudal fin has been described and figured by Hollister (1936). 

Squamation. The structure of the scales has been dealt with by Hollister (1939) 
and Cockerell (1912). Consequently only brief mention of them is made here. 

There are 41-48 lateral line scales (in specimens examined the mode was 44). 
The transverse count in front of the dorsal fin is 5/6. Scales of the lateral line are 
rounded, the transverse and horizontal axes being equal. The anterior border shows 
three scallops delimited by four radii which extend to the centrally placed nucleus. 
The tube containing the lateral line has secondary branches, which do not anastomose 
(cf. Megalops). 

The general body squamation shows a variation in the scale shape. Those scales 
immediately behind the cleithrum are deeper than long ; those upon either side of 
the dorsal mid-line are decidedly longer than deep while those on the caudal 
peduncle are only slightly longer than deep. The scales which sheath the base of 
the anal fin (there are no sheathing scales on the dorsal fin, cf. Elops) are smaller 
than other body scales and are ovoid to irregular in shape. 

Despite differences in shape the scale ornamentation remains constant. The 
embedded portion of the scale is marked by concentric circuli composed of many 
partially fused tubercles. In the exposed portion the circuli break up to an ir- 
regular pattern of tubercles. The posterior margin of the scale is very thin, un- 
marked by circuli or tubercles. Instead the border is finely crenulated. In large 
scales (i.e. those of large individuals) the posterior scale margin is frequently frayed. 

An exception to the above scale types is the scale at the axis of the pelvic and 
pectoral fins. The axillary scale is elongate, two and a half times as long as deep, 
rounded anteriorly, pointed posteriorly. The scale is attached to the body wall by 
a thick flap of skin which inserts on the medial face of the posterior half of the scale. 
Circuli are absent and the radii, which originate from the anteriorly placed nucleus, 



FOSSIL AND RECENT 



63 




a 
3 
o 



a 

o 

u 



T3 
U 



in •— j 

fl 6 

W S 
00 

'c/T aj 

OJ ,g 

C C 



> 

=8 



> 



s 

o 






64 ELOPIFORM FISHES 

are irregular in shape. The thinness of the scale, plus the lack of ornamentation 
(circuli), imparts a flexibility to this structure. All scales show bone cells in the 
embedded portion. 

Genus MEGALOPS Lacepede, 1803 

Diagnosis (emended). Megalopid fishes reaching 1 100 mm in length. Neuro- 
cranial roof markedly convex above the cranial vault. Parietals as broad as long. 
Dilatator fossa broad, shallow and without a roof. Process of intercalar forming the 
lateral wall of the periotic bulla. Jugular canal opening posteriorly into the sub- 
temporal fossa. Maximum depth of the neurocranium occurring at the auto- 
sphenotic level. Maxilla not extending behind the eye. Scales with more than 
five anterior (basal) radii. 

Type and only species. Clupea cyprinoid.es Broussonet. 

Megalops cyprinoides (Broussonet) 
(Text-figs. 30-34) 

Diagnosis (emended). Megalops in which the origin of the pelvic fin lies beneath 
that of the dorsal. Maximum depth of body equal to or greater than the head length. 
Dorsal fin with 17-20 rays, anal 24-27 rays. Vertebral column with 67-68 vertebrae 
of which approximately 30 are caudal. Lateral line with 36-41 scales. Caudal 
fringing fulcra 1-9 in number. 

Habitat. Coastal fishes, sometimes entering freshwater. Indo-Pacific, between 
40 N and 40 S. Longitude extremes, east coast of Africa to Society Islands. 

Remarks. Megalops cyprinoides has often been cited in the literature. The 
following works are considered to be the most important anatomical studies : general 
cranial anatomy, Ridewood (1904) ; intercalar, with respect to the otophysic con- 
nection, Greenwood (1970a) ; branchial arches and associated musculature, Holts- 
voogd (1965) ; ligaments and musculature concerned with the feeding and gill 
ventilation movements, Vrba (1968) ; otophysic connection, de Beaufort (1909) ; 
caudal anatomy, Regan (1910) and Hollister (1939). 

The above literature, plus the preceding description of the closely related Tarpon 
atlanticus, renders an osteological description unnecessary. It remains to point out 
the various differences between the two Recent forms. 

The proportions of the skull roof above the otic region of the skull differ from those 
in Tarpon. Thus in Megalops the width at the occiput is relatively greater and the 
distance from the base of the supraoccipital crest to a line drawn between the auto- 
sphenotic spines is considerably less than half the length of that line. In Tarpon this 
latter ratio is considerably greater than half, and thus the length of the otic region 
in Tarpon is greater. 

The parietals are as broad as they are long and show smooth margins. They are 
thus like the parietals of young Tarpon but unlike the adult Tarpon condition, where 
these bones are approximately three times as long as broad and exhibit zig-zag 
sutures anteriorly and posteriorly. 



FOSSIL AND RECENT 



65 




03 



•s. 






3 

2 





•^H 


r~ 


^ 




u 


E ■ 




ID 


-2 




1 




&e 



o 

6 



66 



ELOPIFORM FISHES 

soc 



exo 



ot. s c 




par 



Fig. 31. Megalops cyprinoides (Broussonet). Neurocranium in posterior view. 
Based on same specimen as Text-fig. 30. 

The neurocranium of Megalops is relatively deeper and the maximum depth occurs 
at the sphenotic level, whereas in Tarpon the maximum depth is at the occiput. The 
depth of the Megalops neurocranium is due to the depth of the prootic bone and 
parasphenoid ascending wings which in turn may be related to the invasion of the 
skull by the swimbladder in the following manner. Megalops has a large eye, 
relatively larger than that in Tarpon, and thus might be expected to have a large 
posterior myodome. The extreme anterior end of the swimbladder diverticulum is 
received in a deep pit within the prootic bone immediately above the prootic bridge, 
which is, in consequence, displaced ventrally. This lowering of the myodome roof 
would restrict its volume. Hence it is suggested that the myodome volume is 
maintained by effectively lowering its floor, i.e. by deepening the prootic and para- 
sphenoid walls. Correlated with this deepening is the angled parasphenoid which 
is in contrast to the relatively straight bone seen in Tarpon. 

The intercalar bone of Megalops has been well described by Greenwood (1970a), 
who recognized three interconnected portions ; a basal portion, a lateral shield-like 
part and a saddle-like wing. The lateral shield-like portion forms a complete bony 
outer wall to the periotic bulla and is joined to the prootic, basioccipital and exoc- 
cipital by a syndesmotic union. The anterior swimbladder diverticulum occupies 



FOSSIL AND RECENT 



67 



all of the periotic bulla, leaving no space for the posterior passage of the jugular vein. 
The jugular vein has therefore to take a different course from that seen in most 
elopiforms. The vein leaves the pars jugularis above the level of the prootic- 
intercalar bridge and runs for a short distance along the floor of the subtemporal 
fossa before passing postero-ventrally through a foramen which represents the space 
left between the prootic-intercalar bridge and the exoccipital (cf. a similar foramen 
in Elops, p. 12), and leaves the cranium by a foramen between the intercalar and 
the exoccipital (Greenwood 1970a: pi. 2, ophv). This condition is rather different 
from that seen in Tarpon which has been described above. The position of the jugu- 
lar vein opening in the prootic of Tarpon is primitive. Posteriorly, the jugular vein 
of Tarpon has a slightly different relationship to the intercalar. The jugular vein 
runs back in a groove which is entirely surrounded by the intercalar, there being no 
contribution by the exoccipital. In short, the jugular vein leaves the cranium of 
Megalops between the exoccipital and the shield-like expansion of the intercalar, 
whereas in Tarpon its exit is surrounded by the basal, crescentic and shield-like parts 
of the intercalar. 



pto 



pmx 



101 



oen 



m s 




20mm 



qu iop sy 



Fig. 32. Megalops cyprinoides (Broussonet). Cranium in left lateral view. 
Based on B.M.N.H. 1876.3. 11. 1. 



68 



ELOPIFORM FISHES 








Fig. 33. Megalops cyprinoides (Broussonet). Diagram to illustrate variations in caudal 
osteology. Neural arches stippled, neural spines black. 



The only other significant cranial difference between the two genera concerns the 
length of the maxilla. The maxilla of Tarpon extends well behind the level of the 
eye. That of Megalops nearly always terminates at a level beneath the posterior 
eye border. In two specimens out of a total of 24 examined (ranging in S.L. from 
48 mm to 365 mm) the maxilla protruded very slightly beyond the posterior orbital 
margin. 

In the postcranial skeleton Megalops differs from Tarpon in certain respects 
which are mentioned in the respective diagnoses. Primarily they concern the posi- 
tion of the ventral fins, the ray counts of dorsal and anal fins and the vertebral and 
lateral line scale counts. 



FOSSIL AND RECENT 



69 



Hollister (1939) added further differences in the morphology of the lateral line 
scales and these views are endorsed by the present author. However the differences 
in the caudal anatomy mentioned by Hollister (1939) are not so consistent as that 
author implied. For instance, Hollister found a caudal fin-ray count of 34-35 in 
Megalops while that in Tarpon never exceeded 32. Such a difference does not seem 
apparent in the specimens examined here. Hollister (1939) also stated that the 
basal tip of the first uroneural extended to the anterior part of the second preural 
centrum in Megalops whereas that of Tarpon never extended beyond the first preural 
centrum, a view endorsed by Greenwood (1970a). While this is true of smaller 
individuals, some large individuals of Tarpon exhibit the same uroneural structure 
as in Megalops. 

The variability of the neural arches and spines associated with the first two preural 
centra of M. cyprinoides has been noted by Greenwood (1970a). To these variations 
may be added two more conditions, the range of such structural variation being 
summed up in Text-fig. 33. The neural arch structure of Tarpon is more constant 




4 mm 




10mm 



B 



Fig. 34. Megalops cyprinoides (Broussonet). Diagram to illustrate fringing fulcra 
(stippled) preceding the uppermost principal ray (arrowed). A, a young individual 
B.M.N. H. 1890.2.26. 187 ; B, an old individual B.M.N. H. 1913.4.7.1. Broken line 
indicates the limit of the body squamation. 



70 ELOPIFORM FISHES 

(Greenwood 1970a) ; the only variation is the occasional double nature of the second 
preural vertebra. 

One of the characteristic, albeit retentional, features of the Megalopidae is the 
presence of fringing fulcra on the dorsal surface of the upper principal caudal fin-ray. 
Regan (1910) noted these in both Megalops and Tarpon. In both genera fringing 
fulcra are a very late development ; in Tarpon fringing fulcra are not apparent before 
the fish has reached 170 mm S.L. whereas in Megalops (which is a much smaller 
species) fringing fulcra do not appear before 116 mm S.L. Tarpon never shows more 
than four fulcra (Regan 1910) and generally only one such element is present. 
Megalops, on the other hand, shows a prolific development of about eight in the largest 
individuals. The fringing fulcra of Tarpon are long and slender, the anterior fulcrum 
(generally the only one) being overlapped to a considerable extent by the posterior 
dorsal accessory ray. Those of Megalops are short, relatively broad and are but little 
overlapped by the preceding ray so that in many specimens examined the fulcra 
are not easily distinguished from the terminal segments of the accessory ray. The 
fringing fulcra are discussed further elsewhere (p. 199). 



Genus PROTARPON gen. nov. 

Diagnosis. Megalopid fishes in which the neurocranial roof is flat, marked by 
prominent ridges. Parietals slightly longer than broad. Dilatator fossa narrow, 
deep and roofed by pterotic and autosphenotic. Posterior opening of the jugular 
canal below the level of the subtemporal fossa. Maximum depth of the neuro- 
cranium at the occiput. Maxilla extending behind the eye. Operculum without 
antero-dorsal indentation. Dentary shallow anteriorly. 

Type-species. Megalops prisons Woodward, 1901. 

Protarpon priscus (Woodward) 
(Text-figs. 35-38) 

1844 Hypsodon toliapicus Agassiz : 5, pt. 1 : 101 (nom. nud.). 

1844 Megalops priscus Agassiz : 5, pt. 2 : 114 {nom. nud.). 

1845 Megalops priscus Agassiz ; Agassiz : 308 {nom. nud.). 
1854 Megalops priscus Agassiz ; Owen : 172 {nom. mid.). 

1 90 1 Megalops priscus Woodward : 24, pi. 3 figs. 3-6. 
1966 Megalops priscus Woodward ; Casier : 118. 

Diagnosis. See Woodward (1901 : 24). 

Holotype. B.M.N.H. 36070, head with opercular apparatus, the skull roof 
crushed laterally. 

Material. The holotype and paratypes B.M.N.H. P.356, P.637, P.4i54a, 
P. 26694 (formerly P. 1698) and P. 26743. With the exception of the holotype and 
P. 26743 these specimens consist of the posterior half of the neurocranium in varying 
states of preservation. The only specimens which can be definitely assigned to P. 
priscus are those with cranial remains. 



FOSSIL AND RECENT 



71 




soc 



epo 



Fig. 35. Protarpon priscus (Woodward). Neurocranium, dorsal view of the otic 
region. Based on B.M.N.H. P.356. 



Formation and locality. London Clay (Ypresian) of Sheppey, Kent, England. 

Description. Neurocranium. The skull roof is flat, marked only by two promi- 
nent ridges, which attain their greatest development at the sphenotic level. The 
mid-line of the skull roof is slightly depressed, an appearance which is enhanced by 
the presence of these ridges. A smaller specimen, B.M.N.H. P.356, shows weak 
ridges. By analogy with Tarpon and Megalops these ridges contained the supra- 
orbital sensory canals. The supraorbital sensory canal is known to pass posteriorly 
into the parietal bones in Megalops and Tarpon and a similar situation existed here 
as evidenced by the pores upon the parietal bones illustrated. In the otic region the 
roof is parallel-sided, but posteriorly it is produced into three processes ; a median 
supraoccipital with large epiotic processes on either side. 

Each frontal has a square posterior margin and in consequence imparts an ' L '- 
shape to the adjacent pterotic. The medially united parietals are slightly longer 
than broad and have a sinuous suture with both pterotics and frontals. The lateral 
ethmoid is similar to that of Tarpon. 

In posterior view the pentagonal supraoccipital bears a prominent supraoccipital 
crest, on either side of which is a slight depression. The epiotic also exhibits a faint 
depression immediately beneath the epiotic process. This depression is continued 
upon the exoccipital. The exoccipital is large and meets its fellow of the opposite 
side above and below the foramen magnum. Laterally the exoccipital forms the 



72 



ELOPIFORM FISHES 

fahm 



fuv 




asp 



Whm 



30mm 



par 



ptfc 



fica 



Fig. 36. Protarpon priscus (Woodward). Neurocranium, right lateral view of the otic 
region. Based on B.M.N. H. P. 356 and P. 26694 (formerly P.1698). 



ventral margin to the post-temporal fossa and at this point the bone bears a promi- 
nent groove which leads towards, but fails to reach, the occipital condyle. Ventrally 
the exoccipitals form with the basioccipital a tripartite suture with the first vertebral 
centrum. This centrum is functionally part of the neurocranium. The centrum 
exhibits a shallow, asymetrically placed notochord pit. The outer surface of the 
centrum is generally smooth save for a few longitudinal grooves on its dorsal half. 
If, as in the Recent megalopids, a neural arch was attached to this element, its union 
with the centrum must have been weak since there is no evidence of articulatory 
facets. 

A lateral view shows that the neurocranium is deepest at the occiput, a resemblance 
to Tarpon. The dilatator fossa is shallow and roofed by the pterotic which forms 
much of its inner wall and floor. The hyomandibular facet slopes ventro-anteriorly, 
there being a slight downward and lateral flexure at the extreme anterior end. 
The facet consists of two oval depressions linked by a somewhat narrower groove. 
Like the dilatator fossa the hyomandibular facet is formed largely by the pterotic. 
The anterior region of both the facet and the fossa is formed by the autosphenotic 
which is produced laterally as a weakly developed autosphenotic spine. 

The subtemporal fossa is large and formed by the pterotic, exoccipital and prootic, 
which line the dorsal, postero-ventral and antero-ventral regions respectively. It 



FOSSIL AND RECENT 



73 



is significant that the anterior border of the subtemporal fossa is ill-defined since this 
invites comparison with Tarpon. The ventral border is clearly demarcated by a 
prominent prootic-intercalar bridge which stands clear of the underlying exoccipital, 
leaving a small foramen between the bridge and the lateral cranial wall. 

Beneath the subtemporal fossa the exoccipital contacts the basioccipital ventrally 
and the prootic anteriorly. At the junction of these three bones there is a shallow 
depression. The depressions of either side are only separated from one another by 
a double layer of bone and consequently the saccular recesses must have lain above 
these depressions, as in Tarpon. The exoccipital is pierced by two foramina. The 
larger of these, the vagus foramen, is situated beneath the intercalar and directed 
postero-ventrally. The smaller glossopharyngeal foramen faces antero-ventrally 
and is situated in front of and somewhat below the level of the vagus foramen. 

The intercalar overlies part of the exoccipital laterally and part of the exoccipital, 
pterotic and epiotic posteriorly. Much of the intercalar is sutured to the exoccipital. 
Anteriorly the intercalar is produced as a thin rod to meet a similar projection from 
the prootic so forming the prootic-intercalar bridge. Posteriorly the intercalar 
bears a small knob, bordered dorsally and ventrally by grooves, the ventral groove 
being particularly prominent. The knob of the intercalar probably provided an 
insertion point for a tendon from the post-temporal. The prootic forms much of the 
lateral wall of the otic region and is pierced by three large foramina. Ventrally 
there is a downwardly directed foramen for the orbital artery, dorsally a large 
upwardly facing hyomandibular foramen, while beneath the prootic region of the 
prootic-intercalar bridge is the posterior opening of the pars jugularis. This latter 



epo pr 



soc 




pto 



par 



f uv 



Fig. 37. Protarpon priscus (Woodward). Neurocranium in posterior view. 
Based on B.M.N. H. P.356. 



74 ELOPIFORM FISHES 

opening is continuous with a short horizontal groove upon the prootic. At the level 
of the foramen for the jugular vein, the medial surface of the prootic is produced 
horizontally to meet its partner of the opposite side in the prootic bridge. This 
bridge is notched anteriorly in the mid-line where it forms the posterior border to 
the pituitary foramen. The abducens nerve pierced the prootic bridge. The an- 
terior, vertical face of the prootic is pierced by a large trigemino-facialis foramen. 
The details of the trigemino-facialis chamber are similar to those described for 
Tarpon. Medial to this foramen lies a small oculomotor foramen. 

The parasphenoid, which is only seen in one specimen (B.M.N.H. P. 26694), 
extends almost to the posterior end of the neurocranium. Throughout its known 
extent (i.e. to mid-orbital level) the bone is relatively narrow. Below the orbit the 
parasphenoid bears a thin row of teeth in the mid-line. The teeth extend posteriorly 
to beneath the ascending wings. Ventrally the posterior end of the parasphenoid 
shows a deep median groove which fades anteriorly. A faint lateral groove is seen 
immediately behind the small ascending wings. The internal carotid foramen lies 
at the base of the ascending wing. 

Hyopalatine bones. Very little is known of this series. The hyomandibular 
articulates with the neurocranium by a single expanded head which is weakly divided 
into anterior and posterior regions. The shaft of the hyomandibular is inclined 
ventrally and slightly posteriorly. Its ventral limit is unknown as it is overlapped 
by the large metapterygoid. The triangular quadrate is slightly longer than deep 
and bears a large condylar surface ventrally. Posteriorly the quadrate is deeply 
grooved and receives the rod-like symplectic. The posterior quadrate border is 
thick and provided an attachment point for the deeper divisions of the adductor 
mandibulae musculature. The preoperculum rests against the posterior margin 
of the quadrate. As in Tarpon there is a small notch upon the quadrate immediately 
behind the articulatory condyle. 

Dermal upper jaw. The upper jaw extends from the snout to a level behind the 
orbit. The oral margin of the jaw is slightly curved, less so than in Tarpon or Mega- 
lops. The premaxilla is relatively small with its oral border equalling one-quarter 
of the total jaw length. The dorsal border of the premaxilla is produced into a 
prominence (it hardly deserves the term articular process) which fits into a slight 
groove upon the head of the maxilla. 

The maxilla is a moderately deep, parallel-sided bone, overlain posteriorly by 
two large supramaxillae which resemble those of Tarpon. The head of the maxilla 
tapers anteriorly before expanding into a dorsal palatine malleolus and an anterior 
ethmoid prominence. 

The oral border of both premaxilla and maxilla bears a thin band of very small 
villiform teeth. Their posterior extent upon the maxilla is unknown. There appear 
to be fewer teeth than in either of the Recent megalopids. 

Mandible. The mandible projects in front of the premaxilla. The oral border 
curves slightly upwards from the shallow dentary symphysis. At one-third of the 
way back the oral border rises more steeply to form a rather elongated coronoid 
process, as in Recent megalopids. The teeth borne along the oral border are similar 
to those on the upper jaw. 



FOSSIL AND RECENT 



75 



pto 



mx 




sop 



Fig. 38. Protarpon priscus (Woodward). Cranium in left lateral view. 
Based on holotype. 



Ventrally the margin of the jaw exhibits a slight inflexion. The surface of the 
bone in this region is marked by coarse rugae. The dentary forms most of the lower 
jaw, but posteriorly the articular is produced into a prominent process. The mandi- 
bular sensory canal pierced the posterior edge of this process and ran through the 
mandible opening by small pores which become more numerous nearer the dentary 
symphysis. 

Circumorbital series. The bones of this series are thin and marked only by faint 
radiating ridges. The first infraorbital is rounded anteriorly and bears a small 
dorsal prominence which lies posterior to part of the trapezoidal antorbital bone. 
The antorbital is considerably larger than, and of a different shape from, that element 
in Tarpon and Megalops. Beneath the orbit the infraorbital bones are narrow but 
the extent of the second cannot be seen. The third, fourth and fifth members of the 
series are broad, extending posteriorly over the preoperculum. 

The infraorbital sensory canal runs within the infraorbitals close to the orbital 
margin, giving off short branches posteriorly. Within infraorbital 1 these branches 
are numerous and directed posteriorly. 

Opercular series. As a consequence of the posterior quadrate/mandibular articula- 
tion the preoperculum is composed essentially of a ' vertical ' limb. The bone is 



7 6 



ELOPIFORM FISHES 



narrow dorsally, but ventrally it widens considerably. The concave anterior margin 
of the preopercular is thickened mid-way along its length to receive part of the super- 
ficial adductor mandibulae musculature. The preopercular sensory canal ran close 
to the anterior margin and opened to the surface of the bone by several large pores. 

The operculum is a large bone showing a rounded dorsal margin. The ventral 
margin is inclined. Like the other opercular bones the surface is marked by ridges 
radiating from the centre of ossification. The suboperculum continues the contour 
of the operculum and is partially overlapped by the latter. An interoperculum is 
present beneath the preoperculum. 

There are at least 20 branchiostegal rays. The posterior members of the series 
are broad and bear the same surface markings as the opercular bones. The anterior 
branchiostegal rays are fine and needle-like and when the mouth was closed the 
anteriormost branchiostegal rays were situated beneath the level of the first infra- 
orbital. 

Between the lower jaw rami and overlying the anterior branchiostegal rays is a 
large gular plate, at least half as long as the mandible. The gular plate is widest 
posteriorly where the width is equal to one-third of its length. 



asp 




pto 



epo 



soc 



30 mm 



Fig. 39. Protarpon cf. prisons. Neurocranium, dorsal view of the otic region. 

From B.M.N. H. P.4153. 



FOSSIL AND RECENT 



77 



SOC 




fahm 



exo 



15mm 



Fig. 40. Protarpon cf. priscus. Neurocranium, posterior view of dorsal third. 
From B.M.N.H. P.4153. 



Appendix to Protarpon priscus 

B.M.N.H. P. 4153 was considered to be Megalops priscus by Woodward (1901 : 
26, pi. 3, fig. 6) and Casier (1966: 118). This specimen consists of the rear part 
of a neurocranium and although incomplete, it differs somewhat from the typical 
P. priscus. In dorsal view not only is the skull roof relatively wider at the sphenotic 
level, but the widest point of the neurocranium occurs towards the posterior end of 
the hyomandibular facet, and not, as in P. priscus, at the sphenotic level. The 
parietals are relatively smaller and exhibit a characteristic ' waisted ' appearance. 
In posterior view the openings of the post-temporal fossae are circular and not 
vertically elongated as in P. priscus. 

Laterally the neurocranium (Woodward 1901 : pi. 3 fig. 6) shows a large, deep 
dilatator fossa containing a large foramen which probably communicated with the 
post-temporal fossa. The hyomandibular facet is divided into anterior and posterior 
oval depressions, linked by an extremely narrow ledge of bone. Finally, the sub- 
temporal fossa appears slightly smaller in area, resembling that of Megalops. 

The points mentioned are at variance with the neurocranial details seen in the 
(admittedly few) specimens of P. priscus and appear sufficient to remove this speci- 
men (P. 4153) from P. priscus. However, apart from the neurocranium nothing 
else is known about this form and in consequence it would not be justifiable to erect 
a new species for it. Pending a better knowledge of the ontogenetic and individual 
variation of P. priscus (which could only be gained by more specimens) and/or more 
complete material of forms like P.4153, it seems best to regard this specimen as a 
form very much like P. priscus but possibly representing a different species. 



78 ELOPIFORM FISHES 

Protarpon oblongus (Woodward) 

1844 Hypsodon oblongus Agassiz : 5, pt. 1 : 101 {nom. nud.). 

1901 Megalops oblongus Woodward : 26. 

1966 Megalops oblongus Woodward ; Casier : 119, pi. 13, fig. 2. 

Diagnosis. See Woodward (1901 : 26). 

Holotype. B.M.N.H. P. 634, an otic region of a neurocranium, rather poorly 
preserved. 

Material. Only the holotype is known. 

Formation and locality. London Clay (Ypresian) of Sheppey, Kent, England. 

Remarks. This species is retained with reservation. Woodward (1901) dis- 
tinguished it on the basis that the distance from the occiput to the level of the 
sphenotic spines is equal to the width at the sphenotics. Although the precise 
distance between the sphenotic spines cannot be measured in B.M.N.H. P. 634 
(the lateral roof margins are broken in the specimen) Woodward's distinguishing 
character remains essentially true. 

In the Recent Tarpon the proportions of the posterior part of the skull roof change 
slightly throughout ontogeny, there being a slight increase in length relative to width. 
However, in P. prisons the occiput-sphenotic distance never exceeds 75 per cent of 
the intersphenotic distance, and is thus very different from that seen in P. oblongus. 

Casier (1966 : 119) cites the more weakly developed frontal depression and the 
convergence of the ridges on the skull roof as further differences from P. prisons, but 
in view of the great variability of these features in the genus Protarpon it seems best 
to ignore them. 

The parietal bones of P. oblongus are only partially preserved, yet enough remains 
to indicate that their length exceeded twice their width, as in Tarpon, whereas the 
parietals of P. prisons are only slightly longer than wide. However during the 
growth of Tarpon the shape of the parietals changes from square to rectangular. 
The lateral otic wall of the neurocranium is very similar to that of P. prisons. 

In view of the ontogenetic changes seen in Tarpon it could be argued that P. 
oblongus is merely an older individual of P. prisons, but the small difference in size 
between P. prisons and P. oblongus could not account for the considerable differences 
in proportions mentioned above. Irrespective of the true specific identity, the 
important fact remains that the otic region is very similar to P. prisons in the roofed 
dilatator fossa, hyomandibular facet, large subtemporal fossa and the position of the 
vagus and jugular foramina. 



Discussion of the genus Protarpon 

Protarpon is clearly a member of the Megalopidae. With these fishes it shares 
such characters as a deep neurocranium with an otic depression comparable to that 
receiving the swimbladder in Recent forms ; very large post-temporal fossae ; 
large, prominent epiotic bones ; narrow, virtually edentulous parasphenoid ; and 
general characters of the ethmoid region (preservation of this region precludes 



FOSSIL AND RECENT 79 

detailed comparison). The jaw structure is similar to that seen in the Recent forms, 
in particular the deep maxilla with the palatine and ethmoid processes situated 
close together. The convex oral margin shows a similar dentition. In the opercular 
series, similarities are seen in the preopercular sensory canal. Further points of 
similarity include primitive features such as small premaxillae, the large gular plate 
and high number of branchiostegal rays. These last points cannot, of course, be 
held to indicate relationship, but merely suggest that Protarpon has not become 
specialized in other directions from Recent megalopids. 

On the other hand, there are several features suggesting that Protarpon should not 
be included in a Recent genus. Protarpon was included by Woodward (1901) in 
the genus Megalops (which included both Recent species) . In recent years, workers 
(Hollister 1939 ; Greenwood 1970a) have qualified Jordan and Evermann's (1896) 
allocation of the two Recent species to separate genera. The basis of this separation 
has been centred upon the swimbladder/ear linkage (Greenwood 1970a) and certain 
other details of neurocranial anatomy (see diagnosis of Recent forms and the dis- 
cussion of M. cyprinoides included here). The Eocene species differ from both 
Recent forms in such details and thus must be accorded generic status. The name 
Protarpon has been chosen in the hope that it conveys the idea of greater anatomical 
similarity to Tarpon than to Megalops. These differences are noted and discussed 
below. 

The neurocranial roof of Protarpon is relatively flat, in contrast to the convex 
form in Recent species. In these latter forms the cranial convexity is due to a 
deepening of the post-temporal fossae, which meet above the endocranium. Such 
a deepening is not apparent in Protarpon and may indicate that the post-temporal 
fossae of either side remained separate from one another. As in the Recent species, 
the post-temporal fossae probably extended forward to the orbitosphenoids, 
since these latter bones exhibit a wide dorsal separation. The autosphenotic spine 
of Protarpon is weakly developed and the dilatator fossa small and roofed 
by the pterotic. This is unlike the condition in extant species where the 
dilatator fossa is deep. The weak autosphenotic spine is probably correlated with 
the flat skull roof. 

In the otic region of the neurocranium there is a single shallow depression formed 
by the exoccipital, basioccipital and prootic. This depression may have contained 
a simple diverticulum of the swimbladder which must have been only slightly more 
complex than that seen in large specimens of Elops lacerta (Greenwood 1970a). 
The swimbladder in Recent megalopids is received in complex depressions in the 
lateral cranial wall. That the swimbladder projection was simple in Protarpon is 
suggested by the primitive position of the jugular canal opening, the position of the 
vagus and glossopharyngeal foramina and the lack of any supporting crista upon the 
basioccipital. In Recent megalopids the swimbladder has pushed so far forward as 
to shift the vagus and glossopharyngeal foramina to a higher and more posterior 
position. In Megalops the jugular canal opening has also shifted. 

The premaxilla of Protarpon is considerably narrower than that of extant megalo- 
pids. The mandible is relatively shallow, without a very large coronoid process, 
and the symphysis does not protrude to any great extent. The dorsal margin of 



8o ELOPIFORM FISHES 

the operculum is smoothly rounded, without the characteristic indentation seen in 
Tarpon and Megalops. 

In all of the above respects Protarpon is more primitive than the Recent mega- 
lopids. It shows no specialization that would exclude it from being an ancestor to 
either or both of the Recent genera. The apparent greater similarity to Tarpon 
rather than Megalops can be attributed to the relative primitive nature of Protarpon 
and Tarpon. These two last genera share such characters as a relatively shallow 
neurocranium, with the maximum depth at the occiput, a straight parasphenoid, the 
primitive position of the jugular foramen and the lack of a bony covering to the 
swimbladder/ear connection. 

Genus PROMEGALOPS Casier, 1966 

1966 Promegalops Casier : 120. 

Diagnosis (emended) . Megalopid fishes in which the neurocranial roof is markedly 
convex above the cranial vault. Parietals slightly longer than broad. Dilatator 
fossa shallow, partially roofed by a very thin ledge formed by the pterotic and auto- 
sphenotic. Maximum depth of the neurocranium at the level of the autosphenotic. 
Parasphenoid bent at the level of the ascending wings. Mandible shallow, coronoid 
process weakly developed. Operculum with antero-dorsal indentation. 

Type-species. Promegalops signeuxae Casier, 1966. 

Promegalops sheppeyensis Casier, 1966 
(Text-figs. 41-43) 

Remarks. Casier (1966) erected the genus Promegalops for two specimens from 
the London Clay (Ypresian) of Sheppey, Kent, England. For each specimen a new 
species was erected, P. signeuxae (the type-species) and P. sheppeyensis (for a speci- 
men formerly included in Megalops prisons, B.M.N.H. P. 9192). The former is 
known by a nearly complete head, the latter from a neurocranium. Both have 
been adequately described by Casier (1966), but apart from referring the genus to 
the Elopidae (including the Megalopidae) he came to no definite conclusions about 
the affinity of this genus. 

The megalopid nature of Promegalops is shown by the deep skull with a lower jaw 
which is obliquely inclined towards the snout ; a dentary symphysis which protrudes 
slightly ; the otic region of the neurocranium which is short and deep ; and by the 
dermethmoid and mesethmoid which have a similar shape to that seen in Recent 
megalopids (Casier 1966 doubts the presence of the postero-ventral projection of the 
ethmoid which occurs in the Recent species) . 

Promegalops differs from the contemporaneous Protarpon in the following important 
respects : skull roof markedly convex above the cranial vault ; autosphenotic spine 
moderately well developed ; dilatator fossa shallow and partially roofed by a very 
thin ledge of the pterotic and autosphenotic ; neurocranium deep, with the maximum 
depth at the level of the autosphenotic ; parasphenoid bent beneath the level of 



FOSSIL AND RECENT 

fa mx 



81 



SO s c 




asp 



Fig. 41. 



soc 



Promegalops sheppeyensis Casier. Neurocranium in dorsal view. 
From holotype. 



the ascending wings ; the quadrate/mandibular articulation lies beneath the middle 
of the orbit (only known in P. signeuxae) ; lower jaw protrudes beyond the pre- 
maxilla (only known in P. signeuxae) ; operculum exhibits an antero-dorsal indenta- 
tion (only known in P. signeuxae) . 

In all these respects Promegalops is more advanced than Protarpon, and is inter- 
mediate between Protarpon and Megalops. 

Promegalops differs from Tarpon in the deep neurocranium, the angled parasphe- 
noid and the general proportions of the otic region of the neurocranium. Such 
features are found in Megalops and are considered to be advanced. Despite the 
close similarity between Promegalops and Recent Megalopidae (in particular Mega- 
lops) there are certain features of the Eocene species which may be considered primi- 
tive or diagnostic. Primitive features include the shallow lower jaw, with a low, 

6 



82 



ELOPIFORM FISHES 

soc 



exo 




fuv 



par 



Fig. 42. Promegalops sheppeyenis Casier. Neurocranium in posterior view. 

From holotype. 



posteriorly placed coronoid process, and the unexpanded post-temporal. Diag- 
nostic features include the wide frontals, small nasals, and the presence of a large 
foramen (or foramina in P. sheppeyensis) at the base of the dilatator fossa. 

In short, Promegalops probably evolved from a form resembling Protarpon and 
could be ancestral to Megalops, from which it is known to differ only in the wide 
frontals, partially roofed dilatator fossa, the single depression in the otic region 
receiving the swimbladder diverticulum, the unexpanded post-temporal and the 
shallow lower jaw. 



Genus ELOPOIDES Wenz, 1965 

Diagnosis (emended). Megalopid fish in which the head is almost as deep as 
long. Neurocranial roof moderately convex above the otic region. Parietals 
slightly broader than long. Dermal cranial bones coarsely ornamented with rugae. 
Orbit large, the diameter exceeding the preorbital distance. Maxilla extending 
behind the eye. Depth of the operculum exceeding twice its maximum width. 
Scales large, with seven to eight anterior (basal) radii. 

Type and only species. Elopoides tomassoni Wenz. 



FOSSIL AND RECENT 



83 



X 

£ 



13 




Oh 

"o 

e 
s 



>> 



rt .2 



be On 



.5 -5 

TO . — I 



a; ^ 

O 
.. Q 






.0 O 



1= 



■*■ 

ft 



84 ELOPIFORM FISHES 

Elopoides tomassoni Wenz 

1965 Elopoides tomassoni Wenz : 4, text-figs. 1, 2, pi. i, figs. A-E. 

Diagnosis. As for genus, only species. 

Holotype. M.H.N. P. 1961-14-1 from the Upper Albian of Vallentigny, France. 

Remarks. No material was examined, but the illustrations given by Wenz 
(1965 : pi. 1, figs A-E) are adequate for the assessment of the systematic position 
of this species. Wenz described the anatomy and included Elopoides in the 
famiry Elopidae (sensu Woodward, 1901), indicating that Elopoides approached 
Holcolepis ( = Osmeroides) lewesiensis. The similarities between these two genera 
noticed by Wenz include the relative proportions of the head and body, the posi- 
tion of the fins, number of vertebrae, position of the quadrate/mandibular articula- 
tion, the dentition and the absence of enlarged scales at the bases of the paired fins. 
With the exception of the last these features are also found in other elopoids included 
by Woodward (1901) in his family Elopidae, and are of doubtful significance in 
establishing relationship. 

Elopoides differs considerably from Osmeroides lewesiensis. Wenz (1965 : 12) 
indicated the following differences : the cranium of Elopoides is short and the dorsal 
profile is steeply inclined, the orbit is large and the preorbital distance is short ; the 
quadrate/mandibular articulation is situated at the posterior margin of the orbit, 
and the shape of the dentary is different ; the preoperculum is of different shape ; 
and the pelvic fins are nearer to the pectorals than in Osmeroides lewesiensis. To 
these differences may be added the shape of the neurocranial roof, the infraorbital 
series and the opercular apparatus. The numerous differences suggest that Elopoides 
should be removed from association with 0. lewesiensis. 

Several of the features in which Elopoides differs from 0. lewesiensis are points of 
similarity with megalopids. Thus the cranium is relatively short with respect to 
the depth. The head of Elopoides is in fact relatively deeper than any other mega- 
lopid. The otic region of the skull roof is comparatively short and exhibits a moderate 
convexity, both features seen in Recent megalopids. The greatest width of the 
cranial roof occurs at the occiput, as in Tarpon, whilst the well-developed epiotic 
process is a typical megalopid feature. The parietals are relatively large and nearly 
square, resembling the Eocene megalopids and the Recent Megalops (Tarpon is 
specialized in possessing elongate parietals). The coarse ornamentation seen upon 
the dermal bones of Elopoides is an unusual feature among megalopids, but is seen in 
restricted areas in Protarpon prisons. The very large pores opening to the supra- 
orbital sensory canal are only found elsewhere among Elopoidei in young specimens 
of Elops and Megalops. 

The dermal jaws show megalopid features in the deep maxilla with a strongly 
convex oral border, the large supramaxillae, and the flattened head of the maxilla 
immediately behind which there is a dome-shaped protuberance, which in Recent 
megalopids articulates with the palatine medially and the first infraorbital laterally. 
The mandible, as in megalopids, exhibits a deep coronoid process and a well-developed 
' retroarticular process ' fitting into a shallow depression developed on the quadrate 



FOSSIL AND RECENT 85 

behind the articulatory condyle. The symphysis of the lower jaw is shallow as in 
the Eocene Protarpon and Promegalops and is unlike the deepened symphysis of 
Recent megalopids. The lower jaw of Elopoides does not protrude beyond the upper 
jaw. Wenz describes and figures an area of very small teeth on the lateral surface 
of the dentary. In this respect Elopoides is very unusual, not only among mega- 
lopids but also among teleosts in general. It may be that this tooth patch represents 
a detached tooth plate. 

In the opercular apparatus, there is both similarity with, and difference from 
other megalopids. The preoperculum is similar in showing very little curvature and 
no ventro-posterior expansion. The operculum, however, is unusual in being con- 
siderably deeper than wide, imparting a narrow appearance to the opercular series 
as a whole. 

Very little is known of the pectoral girdle and thus comparison is difficult. The 
post-temporal, however, appears to be a wide, plate-like bone as in most other 
megalopids. The scales are large, deeply overlapping and marked by fine circuli 
with ornamentation in the form of fine tubercules confined to the exposed field, 
adjacent to a centrally placed nucleus. This form of scale is characteristic of the 
Megalopidae. 

The above facts suggest that Elopoides be placed in the Megalopidae, differing 
from these fishes only in the shape of the operculum. In cranial proportions, rela- 
tive orbital size and scale morphology Elopoides is more reminiscent of Megalops 
than Tarpon. 



Genus SEDENHORSTIA White & Moy-Thomas, 1940 

1863 Microcoelia Marck : 48, non Guenee 1852 (Lepidoptera). 
1940 Sedenhovstia White & Moy-Thomas : 396. 

Diagnosis (emended). Megalopid fishes in which the neurocranial roof is slightly 
convex. Dilatator fossa broad, shallow and without a roof. Maximum depth of 
the neurocranium at the occiput. Upper and lower jaws with a single row of small, 
pointed teeth. Mandible shallow, with a weakly defined, posteriorly situated coro- 
noid process. Small, hook-shaped ossifications lying in the mid-dorsal line between 
occiput and origin of the dorsal fin. Scales small, with four to five anterior radii. 
Exposed field marked by fine granulations arranged in radiating lines. 

Type-species. Microcoelia granulate Marck from the Campanian of Sendenhorst, 
Westphalia, Germany. 

Material. The following specimens in the B.M.N.H. were examined. Seden- 
horstia dayi (Hay) as represented by P.13885, P.13893, P.47513, P.47920-21 (part 
and counterpart), P.13886, all from the Middle Cenomanian of Hajula, Lebanon. 
The last-mentioned specimen had been prepared in acetic acid. Sedenhorstia sp. 
represented by P. 9985 from the Middle Cenomian of Hajula. Sedenhorstia orientalis 
Goody, represented by the holotype P. 9983, an acid-prepared specimen from the 
Middle Cenomanian of Hajula. Sedenhorstia libanica (Woodward) represented by 
P. 4865 from the Middle Cenomanian of Hakel, Lebanon. 



86 ELOPIFORM FISHES 

Remarks. The genus Sedenhorstia contains four species. 5. granulata (Marck) 
from the Campanian of Sendenhorst, S. libanica (Woodward) from the Middle Ceno- 
manian of Hakel, Lebanon, and S. dayi (Hay) and S. orientalis Goody, both from 
the Middle Cenomanian of Hajula, Lebanon. 

The Lebanese species are the best known and have been the subject of a study by 
Goody (1969a) who removed the genus from the Scopelidae (sensu Woodward 1901) 
to the suborder Elopoidei. Goody (1969a) described S. dayi in some detail and 
indicated many similarities with Recent Elops and Megalops, although no indication 
was given as to which of these Recent forms it was more nearly related. Goody 
(1969a) suggested that the genus be placed in a monotypic family, the Sedenhorstii- 
dae, based upon the presence of ossifications within the dorsal median ligament 
between the occiput and the origin of the dorsal fin. The establishment of a separate 
family, equal in status to the Megalopidae and Elopidae, seems unjustified on this 
basis alone, particularly as calcification within the dorsal ligament is not unique to 
Sedenhorstia (Goody 1969a : 20-21). A survey of the characteristics of Sedenhorstia 
indicates affinity with the Megalopidae and it is suggested that the genus be placed 
in this family. The salient points are discussed below. 

The general body form of Sedenhorstia resembles that of the Recent megalopids 
rather than elopids ; the cranium is moderately deep with respect to its length, the 
orbit is large and the mouth is almost superior. The body was probably slightly 
compressed, as evidenced by the straight pleural ribs which nearly encircle the 
abdominal cavity. In megalopids the body is also compressed while in the Elopidae 
it is rounded. The vertebral column is composed of approximately 50 vertebrae of 
which 25 are caudal and is similar to Tarpon in which there are 53-57 vertebrae of 
which 20-24 are caudal. Megalops exhibits a higher count (c. 68) and is thus like 
Elops species in which the count varies from c. 63 in E. machnata to c. 80 in E. saurus. 

The dorsal fin is situated more posteriorly than in Elops or the megalopids and is 
composed of more rays (23). In this respect Sedenhorstia is distinctive. The anal 
fin, however, is decidedly megalopid in being long and composed of at least 22 rays. 
The first anal pterygiophore is also similar to Megalops and Tarpon, exhibiting a close 
approximation with the vertebral column. The typical megalopid feature of an 
elongate terminal dorsal and anal lepidotrich is not seen in Sedenhorstia. The origin 
of the pelvic fin in Sedenhorstia occurs anterior to the level of the dorsal fin origin 
(in S. orientalis the pelvic and dorsal origin are at the same level), and is thus similar 
to Tarpon. 

The caudal skeleton is of the basic elopoid (or leptolepid) type. In possessing a 
half spine upon the second preural centrum Sedenhorstia is primitive (Patterson 
1968b) and resembles Elops rather than Megalops or Tarpon, but some specimens of 
Megalops do show a half spine and perhaps little significance can be attached to this 
feature in this instance. The base of the first uroneural is unforked and no urodermal 
has been found, conditions which are unlike Elops but in accordance with mega- 
lopids. A typical megalopid feature of Sedenhorstia is the retention of fringing fulcra 
on the upper margin of the caudal fin. S. dayi possesses three such elements while 
there are five in S. libanica. In the number of fringing fulcra Sedenhorstia therefore 
resembles Megalops rather than Tarpon in which there are rarely more than two. 



FOSSIL AND RECENT 87 

The neural arches of the first ural and first preural centra are fused in Sedenhorstia, 
producing a single median expansion which Goody (1969a : 19) suggested may be the 
forerunner of a stegural. Megalopids and Elops always show distinct neural arches 
on these two centra, but in older specimens of the Recent genera there is only a 
slight indication of their individuality. It would therefore seem that the con- 
solidation in Sedenhorstia is of minor significance. 

In the cranium there is a general resemblance to megalopids. The skull is 
relatively deep with the quadrate/mandibular articulation beneath the orbit. S. 
orientalis is unusual among the Sedenhorstia species in that the lower jaw articulation 
occurs immediately behind the orbit. In this respect this species shows an inter- 
mediate condition between Tarpon and Elops. The orbit of Sedenhorstia is rela- 
tively large and the long basisphenoid pedicel in S. dayi suggests a deep myodome, as 
in Megalops. The otic region of the neurocranium is relatively short, as in Megalops, 
and the skull roof in this region exhibits a certain convexity suggesting large post- 
temporal fossae. Recent megalopids show an open dilatator fossa which is deep in 
the dorso-ventral plane. Such a fossa, also seen in Sedenhorstia, differs from the 
shallow, roofed fossa in Elops. The Eocene megalopids more nearly resemble Elops 
in this respect. 

The parasphenoid of Sedenhorstia appears to be edentulous. If teeth were present 
they must have been very small and restricted to the mid-line beneath the basi- 
sphenoid pedicel. A nearly edentulous parasphenoid is found in megalopids, but 
not in Elops. 

The upper jaw shows a marked convexity along the oral border and is relatively 
deep, with two very large supramaxillae. Anteriorly the maxillary head shows a 
simple ethmoid projection immediately behind which there is a rounded process 
articulating with the palatine. In all these features there is a close agreement with 
the megalopids. A further megalopid feature of the upper jaw is the extension of 
the premaxilla beyond the dermethmoid. The maxilla of S. orientalis resembles 
that of Elops in its length, extending well behind the orbit. The lower jaw is similar 
to Elops, Protarpon and Promegalops in remaining moderately shallow throughout 
its length, with an ill-defined coronoid process. The Recent megalopids have a 
prominent coronoid process, which is elongated antero-posteriorly. The teeth on 
the dermal jaws are unusual among elopoids in being pointed and set in a single row. 

The circumorbital bones are relatively narrow and the fifth tapers dorsally, as in 
megalopids. The fifth infraorbital is rectangular in Elops. The first infraorbital 
shows a gently rounded dorsal margin, as in megalopids. In Elops the first infra- 
orbital turns upwards to lie alongside the supraorbital. The number of infra- 
orbitals is constant throughout the elopids and megalopids. Goody (1969a : fig. 1) 
shows six infraorbitals (one more than is usual) in S. dayi but this appears to be an 
error. 

In the opercular series the preoperculum is narrow with the dorsal limb inclined 
antero-ventrally (or vertically in S. orientalis), as in megalopids. The preopercular 
sensory canal opens to the surface by a series of large pores. In Elops the upper 
preopercular limb is inclined postero-ventrally and the sensory canal opens by a 
series of short secondary branches, each with a terminal pore. 



88 ELOPIFORM FISHES 

The scales are relatively small, unlike those of megalopids, and are thin and 
cycloid. The presence of bone cells has not yet been demonstrated. The anterior 
field is marked by 4-5 radii resulting in 3-4 scallops along the anterior margin. 
Tarpon usually exhibits 4 radii, while Megalops shows about 6. Elops and Davich- 
thys scales show at least 11 radii. The ornamentation of the exposed field is like the 
Eocene megalopids in showing a fine granulation composed of minute tubercles. 
A point of distinction is that the granulations in Sedenhorstia appear to be arranged 
in radiating lines. The exposed surface of elopid scales is unornamented except for 
the area immediately adjacent to the nucleus. 

It is regrettable that the ethmoid and intercalar of Sedenhorstia are not well known, 
for it is in these areas that the megalopids are most clearly characterized. Never- 
theless, the features referred to above serve to establish that Sedenhorstia exhibited 
a morphology comparable with that of the Megalopidae. Features of the dentition 
and perhaps the peculiar ossifications in the dorsal ligament divorce Sedenhorstia 
from the main megalopid line, but its close alliance with this lineage seems certain. 
The phylogenetic position of Sedenhorstia is further analysed in the general dis- 
cussion of the Megalopidae (p. 201). 



Genus PACHYTHRISSOPS Woodward, 1919 

? 1914 Parathrissops Eastman : 423 (name preoccupied : Parathrissops Sauvage, 1891 : 37). 
1 919 Pachythrissops Woodward : 128. 

Diagnosis. See Woodward (1919). 

Remarks. The genus is briefly discussed following a proposal by Nybelin (1964) 
that at least one of the contained species shows a resemblance to Megalops. 

The genus Pachythrissops is known from three species. The type-species, P. 
laevis Woodward, is from the Upper Portlandian of Dorset, England ; P. vectensis 
Woodward is from the Weald Clay (?L. Aptian) of the Isle of Wight, England, and 
P. propterus (Wagner) is from the Kimmeridgian of Solenhofen, Germany. Question- 
ably, the so-called Parathrissops furcatus Eastman is also placed in this genus. 

Several complete specimens of the type-species were examined together with two 
specimens of P. vectensis and one specimen (B.M. N.H. 37056) which is probably 
a young individual of P. propterus. 

Discussion. During a revision of the genus Thrissops Agassiz, Nybelin (1964) 
briefly described P. propterus and suggested that it was only remotely related to 
Thrissops and Allothrissops. The latter genera are considered to be related to the 
Ichthyodectidae. Furthermore, Nybelin tentatively proposed that P. propterus 
may be an ancestor of Megalops, quoting similarities in premaxillary dentition, 
caudal fin fulcra and other unspecified characters. 

Bardack (1965) included Pachythrissops (recognizing the type-species as P. 
propterus) in the family Chirocentridae (sensu Saint-Seine 1949) and endorsed 
Nybelin's views, quoting the presence of a gular plate in P. propterus as further 
evidence for separating this species from Allothrissops and Thrissops. Bardack 



FOSSIL AND RECENT 89 

(1965 : 34) also stated that P. propterus is genetically distinct from the English 
species. In support of this statement Bardack refers to five points of dissimilarity 
between the English and Solenhofen species : the proportions of the body ; structure 
of the neurocranium ; form of the dentition ; number of vertebrae and the relative 
positions of dorsal and anal fins. I have been unable to examine ' typical ' specimens 
of P. propterus but the description and figures given by Nybelin (1964) are sufficient 
to invalidate the above points. Firstly, the body proportions are very similar in 
both the English and Solenhofen species and differ somewhat from Thrissops and 
Allothrissops. The head length of all Pachythrissops species is contained approxi- 
mately four times in the standard length and thus differs from the contemporaneous 
' chirocentrids ' in which the head length never exceeds one-fifth of the standard 
length. In the body depth there is virtually no difference between P. laevis and P. 
propterus whereas P. vectensis appears slightly more slender. The relative depth of 
the cranium is also similar in all the species of Pachythrissops. 

The unspecified differences in neurocranial structure mentioned by Bardack 
(1965 : 34) are equally difficult to understand. The degree of ' similarity or differ- 
ence ' in neurocranial structure may be taken at various levels depending on how 
well the neurocrania are known. None of the Pachythrissops species is well known 
and in consequence comparison must be at a superficial level. In as much as the 
English and Solenhofen species show shallow neurocrania with a flat roof and a 
straight, edentulous parasphenoid, with a weak development of the lateral ethmoid 
region, there is a degree of resemblance. Such resemblances are of a primitive nature 
with the exception of the weakly developed lateral ethmoid, and are found in other 
contemporaneous teleosts. 

The number of vertebrae is similar in P. laevis and P. propterus ; the count for 
P. vectensis is unknown. P. laevis shows 53-56 vertebrae of which approximately 
22 are caudal. According to Nybelin (1964 : 34) P. propterus has 50-54 vertebrae 
(24-25 caudal). 

The differences in the relative positions of dorsal and anal fins (Bardack 1965) 
appear to be erroneous. In both P. laevis and P. propterus the fins are posteriorly 
placed and the dorsal fin originates slightly anterior to the anal fin. The slight 
difference in the position of the dorsal fin in these two species is of minor importance 
since it is often subject to specific and even ontogenetic variation. 

Further points of similarity between the English and Solenhofen species are seen 
in the jaws. The premaxilla forms nearly one-third of the oral margin of the jaw. 
The dentary exhibits an inflected ventral margin and the oral margin rises steeply to 
produce an elongate coronoid process (this feature is seen in several other contem- 
poraneous teleosts, e.g. Allothrissops and Leptolepis). The dentition is represented 
by a band of small teeth upon the premaxilla, maxilla and dentary. Those of the 
dentary become slightly larger at the symphysis. Saint-Seine (1949) attributed 
large toothed forms to Pachythrissops and concluded that this genus was carnivorous. 
It is probable that the specimens he attributed to Pachythrissops were in fact 
Thrissops formosus. 

The above brief notes suggest that there is, at present, no basis for separating 
Pachythrissops propterus from the English species. 



9 o 



ELOPIFORM FISHES 




sop 



Fig. 44. Pachythrissops laevis Woodward. Cranium in left lateral view. Composite 

of several B.M.N.H. specimens. 



The systematic position of Pachythrissops presents many problems, which stem 
from the extreme primitiveness of this form (a phenomenon not unknown in this 
work). An association of Pachythrissops with Allothrissops and Thrissops seems 
unlikely and the suggestion that Pachythrissops is a stem ichthyodectid (Saint- 
Seine 1949) is untenable. 

Some of the characters which unite Allothrissops and Thrissops with one another 
and with the ichthyodectids include the following : 

i. Body form elongate with a small, deep head and an elongate anal fin 

originating anterior to a short-based dorsal fin. 
ii. Pectoral fin with the outer rays somewhat enlarged and flattened. 
iii. Vertebrae marked with a prominent longitudinal strut. 
iv. Apparent lack of epipleural intermuscular bones, 
v. Dentition in the form of a single row of small (Allothrissops) or prominent 

conical teeth (Thrissops). 
vi. The preopercular sensory canal with many long branches extending postero- 
ventrally in the ventral half of the bone. Although the individual branches 
may not be seen in the Cretaceous ichthyodectids there is a series of many 
pores near the ventral margin of the preoperculum indicating that there are 
long branches from the main preopercular sensory canal, 
vii. A stout basipterygoid process ; this is a retained character. 



FOSSIL AND RECENT 91 

viii. The caudal skeleton contains a series of strap-shaped uroneurals (see below), 
and in the Cretaceous ichthyodectids further specializations occur in the shape 
of the hypurals and the posterior neural and haemal spines (Cavender 1966). 

In features i-v Pachythrissops is more primitive than Allothrissops and Thrissops 
and therefore may not be separated from them on these features. However, the 
last three features indicate that Pachythrissops is more advanced than Thrissops 
and Allothrissops. The absence of a basipterygoid process in Pachythrissops is not 
definitely established, due to imperfections in preservation, but the several neuro- 
crania of P. laevis examined show no evidence of such a process. The preoperculum 
of Pachythrissops differs from Allothrissops and Thrissops in lacking the distinctive 
projection at the postero- ventral angle and the preopercular sensory canal opens 
to the surface by two or three very short branches situated at the angle of the bone. 
Both of these features are common to many lower teleosts and would not invalidate 
the derivation of Pachythrissops from a form like Allothrissops as implied by Bardack 
(1965: fig. 3). 

The caudal skeleton, however, makes it almost impossible to ally Pachythrissops 
with Allothrissops and Thrissops. Pachythrissops laevis shows four elongate 
uroneurals and a series of about three smaller elements situated posteriorly. Nybelin 
(1971) described the caudal skeleton of P. propterus and it agrees with P. laevis. 
In both, there is a full-length neural spine on the second pre-ural centrum and 
four elongate uroneurals. The first uroneural is short and there is a large gap 
between the second and third hypurals. Such an arrangement cannot have given 
rise to that of Allothrissops, etc., and it is highly unlikely that it was derived from the 
latter. To justify this statement a broader view of the teleost uroneural series is 
necessary. 

Patterson (1968a) formulated the evolution of the teleost caudal skeleton from the 
pholidophorid type and described the anatomy of the tail of Leptolepis coryphaenoides 
from the Upper Lias, which may be taken as representing the most primitive teleost 
stage. The definition of a teleost proposed by Patterson (1968a) is accepted and 
would thus include Leptolepis but exclude Pholidophorus. The uroneurals of L. 
coryphaenoides consist of a graded series of six (sometimes seven, Patterson 1968a : 
fig. 9c) elements, the anterior member of this series extending lateral to the second 
or third preural centrum. 

In the upper Jurassic Leptolepis dubia the uroneurals may be divided into two dis- 
tinct series. The anterior series consists of four strap-shaped elements and these 
are succeeded by three short uroneurals, the anterior one of which lies lateral to the 
fourth member of the anterior series. This uroneural arrangement is very charac- 
teristic and may be traced through to Elops. To attain the Elops grade, a fusion 
of the first with the second, and the third with the fourth uroneurals has to be 
postulated. There is both circumstantial and direct evidence for this having hap- 
pened (Patterson 1968a : 226). The third uroneural of Elops corresponds to the 
fifth of Leptolepis dubia. The L. dubia type of uroneural disposition is probably 
basic not only for elopoids, but probably for Clupeomorpha and Protacanthopterygii 
as well, since the two uroneural series are present in the basal members of these 
groups. 



92 



nsppu2 



ELOPIFORM FISHES 

ep1-3 



un 7 




Fig. 45. Pachythrissops laevis Woodward. Caudal skeleton in left lateral view, 
on B.M.N. H. 40333 (holotype) and P.29392. 



Based 



Allothrissops and Thrissops, however, have retained the Leptolepis coryphaenoides 
type. Instead of the fifth uroneural remaining short, this and succeeding uroneurals 
extend basally to lie lateral to the preural and ural centra. Concomitant with this 
development the anterior uroneurals tend to extend forward and in the Cretaceous 
ichthyodectids, which also show at least five elongate uroneurals, the anterior 
members of the series extend as far forwards as the fourth preural centrum. The 
development of this unique uroneural arrangement is probably correlated with 
greater rigidity in the tail. The body form of the ichthyodectids suggests fast 
swimming fishes. The caudal fin is very deeply forked with the caudal fin-rays 
crossing the hypurals at a steep angle and those of the upper lobe lying in line with 
the uroneurals. In the Cretaceous ichthyodectids, and to a lesser degree in Allo- 
thrissops and Thrissops, various degrees of hypurostegy are shown. Hypurostegy 
is by no means confined to the ichthyodectids but is found in several groups of higher 
teleosts, e.g. Scombridae, yet the combination of this feature with the uroneural 
development is unique. 

The caudal skeleton of Pachythrissops is virtually identical with that of Leptolepis 
dubia in showing two distinct series of uroneurals, the anterior group consisting of 
four uroneurals and the posterior of three. To derive an Allothrissops- and Thrissops- 
like caudal skeleton from the Pachythrissops type a simplification of the uroneural 



FOSSIL AND RECENT 93 

series would have first to take place, a view with little to recommend it. Alter- 
natively, the evolution of the Pachythrissops-type from Allothrissops would imply 
that the evolution of two sets of uroneurals occurred twice, independently and at the 
same time. This latter view also seems untenable. Thus Pachythrissops, on caudal 
fin anatomy alone, would appear distinct from the Allothrissops-Thrissops-ich.th.yo- 
dectid lineage. 

Removal of Pachythrissops from association with Allothrissops and Thrissops 
creates the vexatious problem of the true affinity of the genus. The resemblances 
between P. propterus and Megalops mentioned by Nybelin (1964) may also be extended 
to include the English species. Thus the premaxilla bears a band of fine teeth and 
there are fulcra present on the upper caudal lobe. Other similarities may be found 
in the lower jaw, where the oral border is concave and is produced to form an elongate 
coronoid process. Certain resemblances may be seen in the hyopalatine series. The 
hyomandibular and metapterygoid are similar in shape to those of the Megalopidae 
and a prominent metapterygoid-endopterygoid ridge is present in the Recent mega- 
lopids and Pachythrissops. The condyle of the quadrate is succeeded by a shallow 
but well-marked cup which acts as a stop in the downward movement of the lower 
jaw. 

The posterior face of the neurocranium, known only in P. vectensis, exhibits an 
overall similarity with Tarpon. Thus the cranium is deep relative to its width, 
the openings to the post-temporal fossae are large and the epiotic processes well 
developed. In the Eocene and Recent megalopids the first vertebra is incorporated 
with the neurocranium, the exoccipitals and basioccipital forming a tripartite 
surface for it. In Pachythrissops the first vertebra is distinct and the condyle is 
formed solely by the vertebra-like expansion of the basioccipital, as in Albula, 
Pterothrissus and many other 'isospondyls'. In the skull roof certain features agree 
with Tarpon ; the sphenotic spines are prominent and knob-like at their extremities, 
the parietals are longer than broad, and the dilatator fossa appears virtually open 
with only a narrow shelf of bone forming a roof. Conversely, there are several 
differences in the proportions. The skull roof of Pachythrissops is narrow, the parie- 
tals are ornamented posteriorly and several prominent ridges run longitudinally. 
The parasphenoid shows the same form as in Tarpon, and, as far as can be seen, 
is edentulous in contrast to the toothed vomer anteriorly. 

A single specimen of Pachythrissops laevis shows gill-rakers and scattered tooth 
plates which have become detached from the branchial arch elements. In both 
features a resemblance is noted with the megalopids. 

All these similarities are in primitive features and of little use as evidence of 
natural relationship. An undisputed megalopid feature is the modification of the 
intercalar. The intercalar is unmodified in Pachythrissops, yet this does not preclude 
close phylogenetic relationship since it is postulated here that intercalar modification 
was a relatively late (Eocene) development. As is so often the case one is forced to 
rely on general appearance, and there is some agreement in superficial features. 
Thus, Pachythrissops is tentatively placed in the Megalopidae but excluded from 
discussion pending more complete knowledge, not only of the genus itself, but also 
of contemporary teleosts. 



94 ELOPIFORM FISHES 

Suborder ALBULOIDEI Greenwood et al., 1966 

Diagnosis. Elopiform fishes in which the cranium is shallow and broad at the 
occiput. Sensory canals developed as a cavernous system, incompletely covered 
by bone. Ethmoid commissure complete or, more usually, incomplete. Post- 
temporal fossa small, directed antero-medially. Sub-epiotic fossa well developed. 
Dilatator fossa narrow, always with a roof. Autosphenotic spine and epiotic process 
weakly developed. Intercalar small, prootic-intercalar bridge absent. Bulla 
containing the sacculith large. Otophysic connection absent. Lateral ethmoid 
in contact with the parasphenoid. Ectopterygoid process well developed. Quad- 
rate/mandibular articulation beneath orbit or lateral ethmoid. Premaxilla forming 
at least one-third of the upper jaw margin. Mandible with prominent coronoid 
process, shallow symphysis and a strongly inflected ventral margin. Supratemporal 
small, not meeting its partner in the mid-line. Caudal skeleton with six hypurals. 
Scales with dense covering of tubercles in the anterior field. 

Family OSMEROIDIDAE nov. 

Diagnosis. Albuloid fishes in which the snout is not elongated ; the mouth is 
terminal and the ethmoid commissure is complete. Supraorbital sensory canal 
with a branch running into the parietal. Sub-epiotic fossae demarcated medially 
by a pronounced ridge. First vertebral centrum incorporated in the neurocranium. 
Parasphenoid, vomer, dermopalatine, endopterygoid and ectopterygoid with a 
dense covering of small villiform teeth. Basibranchial and basihyal plates with 
similar teeth. Supraorbital large. Posterior infraorbitals broad, completely 
covering the cheek region. Premaxilla without contained sensory canal (in those 
forms in which the premaxilla is known) . Premaxilla and maxilla moving together, 
both elements bearing a band of villiform teeth. Two supramaxillae. Mandible 
with coronoid process situated posteriorly ; articular and endosteal articular ossified 
separately. Branchiostegals more than 14 in number. Gular plate large. Dorsal 
and anal fins short-based. Caudal skeleton in which the second preural neural arch 
bears a half length neural spine. Three uroneurals. Inner caudal rays of each 
lobe with expanded bases which overlap the hypurals. Caudal scute present both 
above and below peduncle. 

Genus OSMEROIDES Agassiz, 1837 

1837 Osmeroides Agassiz : 5, pis. 60b, c. 

1863 Rhabdolepis Marck : 26. 

1868 Holcolepis Marck : 278 (in part). 

Diagnosis. Dermal bones of the cranium stout and heavy, the flat cranial roof 
with a large dermethmoid. Neurocranium with deep dilatator fossa roofed by ptero- 
tic ; intercalar reduced to a knob-like protuberance. Parasphenoid stout, connect- 
ing (sometimes by suture) with the lateral ethmoid, and, like the pterygoid bones, 
bearing a large patch of small, villiform teeth set in shallow sockets. Jaw articula- 
tion beneath the middle of the orbit. Upper jaw with premaxilla forming about 



FOSSIL AND RECENT 95 

half of the functional oral margin. Supratemporal small and triangular, lying behind 
the pterotic. Caudal fin with no gap between the second and third hypurals. 
Body covered by moderately large scales. Scales at the bases of the fins unmodified. 

Type-species. Sal-mo lewesiensis Mantell. 

Remarks. Arambourg (1954) maintained that the correct generic name for this 
form should be Holcolepis. In this he has been followed by Wenz (1965). Aram- 
bourg stated that Osmeroides should be used for the myctophiform fish Sardinioides 
Marck. Goody (1969b) rightly states that the first usage of Osmeroides in a binomen 
was for Osmeroides lewesiensis (Agassiz 1837 : 5, pis. 60b, c) and this precedes the 
first usage of both Holcolepis (Marck 1868 : 278) and Sardinioides (Marck 1858). 



Osmeroides lewesiensis (Mantell) 
(Text-figs. 46-53) 

1822 Salmo lewesiensis Mantell : 235, pi. 33, fig. 12, pi. 34, fig. 3, pi. 40, fig. 1. 

1837 Osmeroides lewesiensis (Mantell) Agassiz : 5 ; pt 1, p. 14, pt 2, p. 105, pi. 60b figs. 1, 2, 
5, 6, 7 (not figs. 3, 4), pi. 60c. 

1838 Osmeroides mantelli Mantell : 307, fig. 1. 



1878 Osmeroides lewesiensis (Mantell) 

1885 Osmeroides lewesiensis (Mantell) 

1888 Osmeroides lewesiensis (Mantell) 

1895 Osmeroides lewesiensis (Mantell) 

1 90 1 Osmeroides lewesiensis (Mantell) 

1964 Osmeroides lewesiensis (Mantell) 



Fritsch : 32, pi. 7, figs. 5, 6, pi. 8, fig. 1. 
Laube : 292, pi. 2, figs. 2, 3 and woodcut. 
Woodward : 322. 
Woodward : 656, pi. 42. 
Woodward : n, fig. 2. 
Danil'chenko : 398, fig. 104. 



Diagnosis (emended). Osmeroides attaining 500 mm S.L. Length of cranium 
equal to twice the maximum width. Parietals longer than broad. Dermal 
skull bones showing prominent ornamentation in the form of coarse radiating 
rugae. Large supraorbital firmly united with the frontal. Parasphenoid extending 
to the rear of the neurocranium. Complete, ossified interorbital septum. Mandi- 
bular articulation at mid-orbital level. Branchiostegal rays 18-21 in number, 
the upper members of the series showing coarse ornamentation. Dorsal fin with 
18-20 rays, arising midway between snout and caudal peduncle. Vertebrae as 
long as deep, marked by irregular longitudinal ridges. Caudal skeleton with four 
uroneurals and three epurals. Scales large, the posterior field with faint tubercles 
arranged in radiating rows, anterior field showing about five prominent radii. 
15-17 scales in the transverse series immediately anterior to the dorsal fin. 

Holotype. B.M.N.H. 4294, a specimen showing the head and abdominal region 
from the Upper Cenomanian (Holaster subglobosus zone) of Lewes, Sussex, England. 

Material. Twenty-five specimens in the B.M.N.H. were examined. B.M.N.H. 
P. 6456 was prepared in acetic acid and B.M.N.H. 49894 was prepared mechanically. 
Specimens come from Upper Albian-Coniacian of S.E. England. The Upper 
Albian record is based on a very imperfect specimen (B.M.N.H. P. 8949) from the 
Gault Clay at Folkestone, Kent. The opercular bones of this specimen are less 
ornamented than the Chalk specimens but otherwise there is close agreement. 



96 ELOPIFORM FISHES 

Description. The description of the neurocranium is based mainly upon the 
acid prepared specimen. Several points were checked against a similarly prepared 
specimen of 0. levis Woodward, B.M.N.H. P. 36240. The differences between these 
two species are noted in the remarks on 0. levis. 

The head plus opercular apparatus is equal to 25 per cent of the standard length. 
The maximum depth of the head occurs at the level of the quadrate/mandibular 
articulation and is equal to 60 per cent of the head length. The diameter of the orbit 
is slightly less than the preorbital distance and represents 20 per cent of the head 
length. The bones of the cranium are thick and the dermal elements are coarsely 
ornamented. 

Neurocranium. The neurocranium is long and shallow, with the cranial vault 
occupying the posterior third of its length. The greatest width occurs at the 
occiput and is equal to 50 per cent of the total neurocranial length. In lateral 
profile the roof is comparatively straight. 

Anteriorly the roof of the cranium is formed by a large dermethmoid which, 
unlike the other roofing bones, is unornamented and smooth. The anterior portion 
of the dermethmoid is diamond-shaped, with the two antero-lateral margins in close 
juxtaposition with the premaxillae. In some specimens there are two to four pores 
aligned transversely across the anterior region of the dermethmoid. This line of 
pores is interpreted as evidence of a bone-enclosed ethmoid commissure. Pos- 
teriorly the dermethmoid extends backwards to the level of the lateral ethmoid and 
is in syndesmotic union with the frontals, the suture line being ' W '-shaped. 

Each frontal is narrow anteriorly but expands above the posterior third of the 
orbit. Much of the cranial vault is roofed by the frontal which is in contact with the 
parietal posteriorly and the pterotic posteriorly and laterally. The frontal of either 
side is united with its partner by a sinuous suture, the meanderings of which increase 
posteriorly. The path taken by the supraorbital sensory canal within the bone is 
marked externally by a strong ridge running the length of the orbital section, but 
posteriorly the ornamentation obscures it. Ornamentation on the frontal (as with 
the other roofing bones) takes the form of irregular tubercles arranged in sinuous, 
sometimes anastomosing lines, radiating from the centre of ossification. The 
ornamentation becomes less conspicuous anteriorly. 

The parietal, which joins its partner in the mid-line throughout its length, is 
rectangular, with the shorter axis transverse. The posterior limit of the parietals 
is such that the supraoccipital is excluded from the skull roof. Ornamentation is 
similar to that on the posterior region of the frontal. 

The pterotic has three external faces ; dorsal (horizontal), lateral and posterior 
(both vertical). The dorsal face, which represents the dermal component, shows 
heavy ornamentation. The roof of both the dilatator fossa and post-temporal fossa 
is formed by this face. A line of pores running along the lateral edge of the dorsal 
face marks the path of the otic division of the cephalic sensory canal system. Pos- 
teriorly the sensory canal leaves the bone by a single large pore. 

The lateral face of the pterotic forms the posterior two-thirds of the hyomandibular 
facet, part of the inner wall of the dilatator fossa and the dorsal region of the sub- 
temporal fossa. Medially this part of the pterotic forms the lateral wall and floor 



FOSSIL AND RECENT 



97 



e com 




ot sc 



pto 



epo 



soc 



Fig. 46. Osmeroides lewesiensis (Mantell). Neurocranium in dorsal view. Composite 

of several B.M.N.H. specimens. 



9 8 



ELOPIFORM FISHES 

SOC 




exo 



10 mm 



f uv 



Fig. 47. Osmeroides lewesiensis (Mantell). Neurocranium in posterior view. 
Based on B.M.N.H. P.6456. 



of the post-temporal fossa. The horizontal semicircular canal within the pterotic 
is indicated externally as an arch-shaped swelling immediately beneath the hyo- 
mandibular facet. 

In posterior view the pterotic is seen to form part of the roof and the entire lateral 
border to the opening to the post-temporal fossa. 

The epiotic forms part of the dorsal and the entire medial margin of the opening 
of the post-temporal fossa. Internally the epiotic forms a small portion of the medial 
wall of the fossa. Dorsally the epiotic is produced as a rounded process which re- 
ceives the upper limb of the post-temporal. The posterior surface of the epiotic 
is deeply excavated, the excavation being continuous with that on the exoccipital 
and together forming a sub-epiotic fossa. 

The supraoccipital is pentagonal in posterior view. Dorsally this bone separates 
the epiotics while ventrally it partially separates the exoccipitals. There is a very 
small supraoccipital crest. Internally the supraoccipital is continued forward as 
two antero-lateral wings which form much of the medial wall of the post-temporal 
fossa of either side. 

The exoccipital meets its partner in the mid-line above and below the foramen 
magnum. Adjacent to the foramen magnum the exoccipital is produced posteriorly 
into a structure resembling a neural arch. The lateral face of the exoccipital forms 
part of the subtemporal fossa, the ventral edge of which is defined by a weak ridge 
on the exoccipital. Posteriorly, beneath the level of the intercalar, the exoccipital 
is pierced by the large, posteriorly directed vagus foramen. Smaller foramina 



FOSSIL AND RECENT 99 

adjacent to the vagus foramen may mark the point of exit of blood vessels and 
minor branches of the glossopharyngeal nerve. The main glossopharyngeal trunk 
left the cranial cavity through a forwardly directed foramen in the exoccipital, 
antero-ventral to the vagus foramen. 

The intercalar forms a cap over the triradiate suture between the pterotic, epiotic 
and exoccipital. The intercalar is small but is produced posteriorly as a prominent 
knob for the attachment of the ventral limb of the post-temporal. 

The basioccipital bears a facet on its posterior face which together with a facet 
on each exoccipital provides an attachment area for the first centrum ; this latter 
structure is functionally part of the neurocranium. Ventrally the basioccipital is 
grooved, but much of this groove is overlain by the posterior part of the para- 
sphenoid. In transverse section the basioccipital is ' W -shaped, the wings of the 
' W ' enclosing the saccular recess of either side. The posterior region of the myodome 
roof is formed by the basioccipital. Laterally the wall of the basioccipital, together 
with adjacent regions of the exoccipital and prootic, is slightly inflated. This 
inflation, developed to a greater degree in 0. levis, represents the lateral wall of the 
otolith chamber. 

The prootic is the largest component of the lateral neurocranial wall. As seen in 
lateral view the bone is synchondrally united with the pterotic and autosphenotic 
dorsally and the exoccipital and basioccipital posteriorly, and syndesmotically 
joined with the parasphenoid ventrally. The anterior region of the subtemporal 
fossa and hyomandibular facet is formed by the prootic. A faint ridge upon the 
prootic continues that upon the exoccipital to form a ventral border to the sub- 
temporal fossa. Beneath this ridge and close to the posterior margin of the prootic 
is a large jugular foramen which leads anteriorly to the pars jugularis. Posterior 
to the jugular foramen the path taken by the head vein is seen as a faint groove 
beneath the prootic-exoccipital ridge. A large hyomandibular foramen is situated 
antero-dorsally to the jugular foramen. A short canal links this foramen with the 
pars jugularis. The orbital artery entered the antero-ventral corner of the prootic, 
close to its union with the parasphenoid ascending wing, and passed upwards 
to pierce the floor of the pars jugularis. The external surface of the prootic is marked 
by a narrow ridge of bone running from above the orbital artery foramen to the level 
of the pars jugularis. This ridge, which represents a site of branchial muscle origin, 
is characteristic of many albuloids. 

The anterior face of the prootic forms part of the posterior orbital wall. Dorsally 
this face contacts the autosphenotic, medially the pterosphenoid and ventrally the 
basisphenoid and parasphenoid. Two foramina pierce the anterior surface. The 
larger and more lateral is the anterior opening of the pars jugularis. The smaller 
foramen opens directly from the endocranial cavity and is the oculomotor foramen. 
From the dorsal margin of the anterior opening of the pars jugularis there is a groove 
running across the surface of the prootic and pterosphenoid. Within this groove 
lay the superficial ophthalmic branches of V and VII. 

The medial wall of the pars jugularis is pierced by two foramina. The posterior 
of these is the facial foramen, and through this passed the hyomandibular trunk. 
The otic, superficial ophthalmic and buccal branches left the cranial cavity through 



ELOPIFORM FISHES 




FOSSIL AND RECENT 101 

the trigeminal foramen. The latter foramen is situated anterior to the facial foramen 
and separated from it by a narrow prefacial commissure. The absence of a separate 
foramen for the profundus ciliaris leads one to suppose that it left the cranium with 
the trigeminal trunk. The medial surface of the prootic is marked by two shallow 
cups, adjacent to each of the foramina described above. These cups housed their 
respective ganglia and they are the only evidence of a defined pars ganglionaris. 

Beneath the level of the pars jugularis the prootics are produced medially as a 
horizontal platform, the prootic bridge. The abducens nerve passed through the 
prootic bridge to the myodome. 

The autosphenotic is tetrahedral in shape, the apex directed laterally and somewhat 
ventrally as a weakly developed autosphenotic spine. The otic ramus of VII 
pierced the anterior face of the autosphenotic. The posterior face of this bone forms 
the anterior wall of the deep dilatator fossa. 

The pterosphenoid is of irregular shape. It contacts the basisphenoid ventrally 
and excludes the prootic from the optic foramen. Of the two foramina piercing the 
pterosphenoid beneath the superficial ophthalmic groove the anterior may be identi- 
fied as having transmitted the trochlearis. The posterior is probably the point 
where the anterior cerebral vein left the cranial cavity. 

The basisphenoid, when viewed anteriorly, is ' Y ' shaped, the wings of the ' Y ' 
contacting the prootic and pterosphenoid while the stem passes antero-ventrally, 
expanding in the sagittal plane as it does so. The ventral contact with the para- 
sphenoid was probably by cartilage. From near the top of the basisphenoid stem 
a short prong of bone on either side passes ventro-laterally towards the myodome 
opening. The significance of this structure is unknown, but it appears to be part 
of the basisphenoid. It is also found in Osmeroides levis, 0. latifrons and in Tarpon 

(P- 50). 

In the dorsal half of the orbit the orbitosphenoid contacts the frontals dorsally 
and the pterosphenoids posteriorly. Although clearly paired dorsally, the orbito- 
sphenoids unite ventrally where they pass into a completely ossified interorbital 
septum. The septum contacts the parasphenoid ventrally and the basisphenoid 
posteriorly. The olfactory tracts passed out of the orbitosphenoid anteriorly. 

The lateral ethmoid forms the anterior wall of the orbit and the posterior and 
outer wall of the nasal capsule. It is composed of transverse and lateral faces which 
meet one another in a sharp ridge. This ridge is expanded laterally as a short spine. 
Beneath the spine the lateral ethmoid is sutured with the expanded parasphenoid. 
The transverse face of the lateral ethmoid meets the frontal dorsally. The un- 
finished nature of the anterior margin of the lateral face suggests that the bone passed 
into cartilage. Much of the ethmoid must have been cartilaginous, but a small area 
of spongy bone lying above the anterior part of the vomer represents a restricted 
mesethmoid. 

»The parasphenoid is stout and extends below most of the neurocranium. Anteriorly 
it widens in the region of the lateral ethmoid but posteriorly beneath the otic capsule 
it becomes narrower. The ascending wings of the parasphenoid are weakly developed. 
The internal carotid foramen is at the base of the ascending wing. Beneath the 
orbit the parasphenoid bears a large tooth plate which bifurcates anteriorly. The 



ELOPIFORM FISHES 



par 



grY&SH 



5Hot 



3ZEhm 



pro 




fahm 



foa 



exo 



fuv 



boc 



Fig. 49. Osmeroides lewesiensis (Mantell). Neurocranium in ventral view. 
Based on B.M.N. H. P.6456. 



FOSSIL AND RECENT 
EHhm 



103 




dpal 



eart 



10mm 



Fig. 50. Osmeroides lewesiensis (Mantell). Hyopalatine bones, opercular series (stippled) 
and mandible of the left side in medial view. Composite of several B.M.N.H. specimens. 



plate bears numerous small villiform teeth of regular size. The vomer bears a large 
tooth plate which is partially divided into lateral halves ; the teeth are similar to 
those on the parasphenoid. 

Hyopalatine bones. The hyopalatine bones are stout and firmly united to one 
another, forming a rigid roof to the buccal cavity. 

The hyomandibular is broad and articulates with the neurocranium by a large 
head which is separable into swollen anterior and posterior regions. The lateral 
surface is marked by two ridges which originate from the articulatory regions and 
converge ventrally to form a single stout ridge running the length of the hyomandi- 
bular shaft. The medial surface of the hyomandibular is marked by a prominent 
groove which leads ventrally to a large foramen. This in turn leads to a canal 
opening beneath the opercular process on the posterior edge. The groove, foramen 
and canal mark the path of the hyomandibular trunk of VII and the efferent hyoidean 
artery. Anteriorly the head of the hyomandibular is produced as a thin wing which 
is partially covered ventrally by the metapterygoid. 

The metapterygoid is large and irregularly shaped. Posteriorly and ventrally it 
lies in the vertical plane but antero-dorsally it twists horizontally to lie in con- 
tinuity with the endopterygoid. Contact with the quadrate was mediated by 
cartilage, but with other bones union was syndesmotic. Dorsally, where it overlaps 
the hyomandibular, a cup-shaped depression is formed which received part of the 
levator arcus palatini muscle. 

The quadrate is fan-shaped and has a deep ' V '-shaped notch which received the 
pointed end of the symplectic. The anterior border of the quadrate is thin and fits 



io 4 ELOPIFORM FISHES 

into a groove along the posterior margin of the ectopterygoid. This latter element 
is boomerang-shaped, with unequal limbs directed postero-ventrally and antero- 
dorsally. The anterior tip of the ectopterygoid fits into a small recess on the pos- 
terior aspect of the autopalatine. Small villiform and rounded teeth are borne along 
the entire buccal surface of the ectopterygoid. The anterior teeth are comparable 
in size to those on the parasphenoid, but posteriorly they become smaller. 

The endopterygoid is large. Its convex buccal surface is covered with many small 
teeth. Posteriorly it is overlapped by the metapterygoid, anteriorly it contacts the 
autopalatine, while laterally it sutures with the ectopterygoid. 

The autopalatine ends anteriorly in a transversely expanded condyle. This 
condyle contacts a process upon the maxilla laterally and presumably lay against 
the cartilaginous ethmoid medially. The dermopalatine is represented by a large 
oval tooth plate which bears similar teeth to those on the endopterygoid and anterior 
region of the ectopterygoid. 

Dermal upper jaw. The upper jaw extends from the snout to the level of the 
quadrate/mandibular articulation. The ventral margin is convex. All bones are 
heavily ornamented with coarse rugae. 

The premaxilla is tapered both anteriorly where it fits against the dermethmoid 
and posteriorly where it fits into a notch on the lateral face of the maxillary head. 
The premaxillae just meet in the mid-line, and form about 38 per cent of the upper 
jaw border. A broad band of many small villiform teeth is present on the pre- 
maxilla. 

The maxilla ends anteriorly in a simple point where it lies against the ethmoid. 
A little further back the dorsal surface is raised in a small knob which articulated 
with the palatine head, probably through a biconcave sliver of cartilage, as in elopoids. 
The posterior end of the maxilla becomes slightly deeper where it is partially overlain 
by the supramaxillae. Teeth similar to those on the premaxilla occur along the 
entire oral margin of the maxilla. There are two supramaxillae set in tandem. The 
anterior supramaxilla is oval and overlain along much of its dorsal margin by a 
projection of the posterior element. 

Mandible. The mandible is slightly longer than the upper jaw. The greatest 
depth of the mandible occurs at the coronoid process, situated at two-thirds of the 
mandibular length from the symphysis. The dentary forms the anterior two-thirds 
of the mandible. From the inturned symphysis the bone expands vertically and 
bears a narrow horizontal ledge along the ventral margin. Beneath this ledge is a 
line of 20-25 small pores set at regular intervals along the entire length of the dentary. 
The mandibular sensory canal ran within the articular posteriorly and the dentary 
anteriorly and communicated with the exterior via these pores. The posterior 
margin of the dentary has the shape of a lazy ' V ', the apex directed towards the 
symphysis. Medially the articular partially overlapped the dentary, there being 
a gap between these elements, the Meckelian fossa, which received the anterior 
extension of the A w division of the adductor mandibulae. The oral surface of the 
dentary bears a broad band of villiform teeth. 

The articular forms the posterior half of the coronoid process and the extreme 
outer edge of the articulatory cup. Ornamentation on the articular is confined to 



FOSSIL AND RECENT 

ace 



105 



pee 




vhh 



10mm 

Fig. 51. Osmeroides lewesiensis (Mantell). Hyoid bar and branchiostegal rays of 
left side in lateral view. From B.M.N. H. P. 5680. 



the postero-lateral aspect. Upon the medial surface the articular bears a prominent 
ledge which is continued for a short distance on the dentary. The dorsal surface 
of this ledge is grooved and provided an attachment area for muscle fascia. In 
0. levis there is a small sesamoid articular resting on the ledge of the articular ; such 
an element was not seen in 0. lewesiensis, but this is probably due to imperfection 
of the specimens. Much of the articulatory cup is formed by an endosteal articular 
which extends forward as a narrow splint overlying the articular ledge. 

Circumorbital series. There are eight circumorbitals of which one, the supra- 
orbital, is anamestic. The supraorbital is a long, narrow bone which lies, for the 
greater part of its length, in a narrow rebate on the frontal. The surface is orna- 
mented in the same fashion as the frontal. The ventral end of the supraorbital 
contacts the small triangular antorbital. 

The first infraorbital (lachrymal) is large with a rounded anterior border. The 
ventral margin overlaps the dorsal extent of the maxilla while dorsally there is a 
small rounded projection. The second infraorbital expands posteriorly where it 
meets the third in an oblique suture. The third, fourth and fifth infraorbitals are 
all wider than deep. A dermosphenotic was not seen in 0. lewesiensis, but the gap 
left between the last infraorbital and the frontal suggests that one was present. 
No definite evidence of rostral ossicles could be found in any of the specimens 
examined. 

Hyoid arch, branchiostegals and gular plate. Both the anterior and posterior 
ceratohyals are stout, the former being three times as long as the latter. The 
posterior ceratohyal shows a notch postero-dorsally which received the ventral end 



106 ELOPIFORM FISHES 

of the interhyal. The anterior ceratohyal is somewhat constricted mid-way along 
its length. Anteriorly the bone expands vertically and transversely. Of the two 
hypohyals, the lower is slightly the larger. The afferent hyoidean artery ran through 
both hypohyals and continued upon the lateral surface of the anterior ceratohyal 
in a prominent groove. 

About 20 branchiostegals (range 18-21) are borne by the ceratohyal. Of these 
the first five or six (the first is taken as the upper or most posterior element) are 
broad and inserted upon the posterior ceratohyal. The anterior rays are slender 
and articulate with the outer face of the anterior ceratohyal. The gular plate is 
large, equal in length to two-thirds of the total mandibular length. 

Cephalic sensory canal system. The supraorbital, otic, mandibular and ethmoid 
commissure sensory canals have been mentioned in connection with the bones through 
which they pass. These and the other cephalic sensory canals run deep within the 
cranial bones and the pores by which they open to the surface are often obscured by 
heavy ornamentation. The sensory canals do not show secondary branches, a 
point of distinction from most elopoids. 

The infraorbital sensory canal ran close to the orbital margin of the infraorbitals. 
Anteriorly the canal terminates by a single pore at the antero-dorsal convexity of 
the lachrymal. A short, narrow antorbital canal runs through that bone and appears 
to end blindly. The only pores seen were those which opened within the lachrymal. 

The preopercular sensory canal ran close to the anterior margin of the preoper- 
culum, diverging slightly from its parallel course at the angle. The canal opened 
to the surface by 9- 11 pores situated below the angle of the canal. 

Within the supratemporal there is a canal which runs at the anterior margin and 
represents part of the supratemporal commissure. Since the supratemporals are 
widely separated from one another the central portion of the commissure must have 
lain in the dermis as in Recent albuloids. Laterally the commissure opens to the 
surface of the supratemporal by a single large pore. A canal linking the cephalic 
sensory system with the lateral line could not be seen. 

Opercular series. All bones, except the interoperculum, are thick and ornamented 
by radiating rugae. 

The operculum is trapezoidal in shape, with the longest edge situated anteriorly. 
The articulatory facet which fits the opercular process of the hyomandibular is 
near the top of this anterior edge. The facet is supported by a horizontal flange 
which also served as a point of insertion for the levator operculi musculature. Be- 
neath this horizontal flange there is an upwardly directed foramen, but whether it 
penetrates the opercular bone (as does a similar foramen in Scomber, Allis 1903) is 
not known. 

The suboperculum is large and continues the posterior border of the operculum. 
Anteriorly the suboperculum sends up a dorsally directed wing under the anterior 
margin of the operculum. 

The interoperculum has the form of an equilateral triangle, with the anterior point 
lying adjacent to the articular, with which it was no doubt connected in life by liga- 
ments. The interoperculum is distinctively short but deep. Laterally the inter- 
operculum is largely overlain by the preoperculum. 



FOSSIL AND RECENT 



107 




Fig. 52. Osmeroides lewesiensis (Mantell). Cranium in left lateral view. 
Composite of several B.M.N.H. specimens. 



The preoperculum has two limbs, horizontal and vertical, the latter being the longer 
and lying closely against the hyomandibular. The anterior margin of the pre- 
operculum is thickened, particularly in the angle, and forms, together with the 
hyomandibular, a prominent ridge for the origin of the adductor mandibulae. The 
preoperculum is relatively narrow throughout its length. 

Pectoral girdle and fin. The supratemporal is triangular, ornamented with coarse 
rugae and lies tightly against the posterior margin of the pterotic. The supra- 
temporal is separated from its partner as in all albuloids. 

The post-temporal is relatively small, consisting essentially of two unequal limbs. 
The upper limb projects forward to overlie the epiotic process, to which it was no 
doubt connected by ligaments. The lower limb is very much smaller, both in length 
and diameter, and projects antero-ventrally to lie against the intercalar. Posteriorly 
the post-temporal overlies the top of the supracleithrum. The lateral line canal 
ran through the post-temporal and continued within the dorsal part of the supra- 
cleithrum. This latter element curves ventrally and overlies the dorsal part of the 
cleithrum. Dorsally and ventrally the supracleithrum ends in rounded margins. 

The cleithrum is by far the largest element in the girdle, extending from the 
posterior angle of the operculum to the medial surface of the interoperculum. The 
anterior margin of the cleithrum is turned medially and thus presents a broad face 
for the attachment of the sternohyoideus musculature. The posterior margin of 
the cleithrum curves postero-ventrally to just above the fin insertion. At this 



io8 ELOPIFORM FISHES 

point the border turns sharply antero-ventrally. Anteriorly and ventrally the tip 
of the cleithrum is inwardly curved and meets its partner in the mid-line. 

Specimen B.M.N.H. P. 6456 shows fragmentary remains of the endochondral 
elements. The scapula fails to enclose the scapular foramen, the outer (upper) 
margin of this foramen being formed by the cleithrum. A notch in the posterior 
border of the scapula marks the position of insertion of the outermost fin-ray. The 
coracoid is incompletely known. Posteriorly this element appears to contact the 
scapula laterally and mesocoracoid dorsally. Anteriorly it is in contact with the 
cleithrum over a very small area. Much of the coracoid is separated from the cleith- 
rum by a large fenestra. The mesocoracoid forms the mesocoracoid arch. The 
bone is narrow in its centre, but dorsally it expands to spread over the medial surface 
of the cleithrum, while ventrally it was in synchondral union with the scapula and 
coracoid. 

There are approximately 15 pectoral fin-rays (as with other fins the exact number 
cannot be ascertained) of which the outermost is distinctive in having a swollen base. 
All rays are articulated and, except the first, are branched. A pectoral splint is 
present. 

Pelvic girdle and fin. The pelvic fin, which originates mid- way between the snout 
and caudal peduncle, is supported by a simple pelvic bone. Each pelvic bone is 
triangular in shape and thickened posteriorly. The surface marking of the posterior 
face suggests that it was covered by cartilage. The number and size of the pelvic 
fin-rays is not known, Woodward (1908) stated that n rays were present but this 
could not be verified. 

Vertebral column. Virtually nothing is known of the vertebral column. Wood- 
ward (1901) quotes a rather wide range of 50-70 vertebrae, and this is all that may be 
suggested with known material. The majority of the centra are as deep as long and 
are marked laterally by a few longitudinal grooves. Exceptions to this general form 
are the last few centra which are slightly longer than deep. 

The neural and haemal spines are stout and the arches are lodged in depressions 
upon the centra. The abdominal centra bear autogenous parapophyses with which 
the pleural ribs articulate. Dorsal (i.e. epineural) intermuscular bones are present 
in the abdominal but apparently not in the caudal region. The absence of inter- 
musculars in the caudal region may be apparent rather than real since in many 
Recent teleosts the posterior intermusculars have no firm attachments to the ver- 
tebral column and are often dislodged when the flesh is removed. 

Modification at the posterior end of the column in support of the caudal fin-rays 
is dealt with in the description of the caudal fin. 

Median fins. The origin of the dorsal fin occurs mid-way between the snout and 
the caudal peduncle. The base length of the fin is slightly greater than its height. 
Woodward (1901 : fig. 2) illustrates 22 fin-rays but the specimens show a lower 
count, 18-19, °f which the first three are unbranched and the first is not jointed. 
The fifth ray is the longest of the series. 

The anal fin originates nearer to the caudal peduncle than to the pelvic fins. The 
base length is almost equal to the length of the longest ray which is the first branched 
and the fourth in a series of 14. The anterior two rays are not jointed. 



FOSSIL AND RECENT 109 

Very few specimens of 0. lewesiensis show the dermal rays of the caudal fin ; 
none shows a complete series and thus the number, nature, length and distribution 
are unknown. The following description of the caudal skeleton is based on B.M.N.H. 
49894. 

The fin-rays are supported by structures associated with four preural and two 
ural centra. The latter are distinguished from those anteriorly by being unorna- 
mented, upturned and longer than they are deep. 

The haemal arches borne by the first four preural centra are firmly united, although 
not fused, with the centra. The haemal spines remain broad throughout their length 
and are thickened along their posterior margins. The anterior margin is produced 
as a thin wing which lies against the preceding spine. Near to the base of the spine 
the anterior margin is notched for the passage of the segmental artery. 

The third and fourth preural centra bear full length neural spines which converge 
distally. The second preural centrum bears a neural arch and a half length spine, 
while the first preural bears only a neural arch, the outline of which is difficult to see. 
All neural ar.m elements are autogenous. 

The hypural series forms a complete fan. Two lower hypurals are borne by the 
first ural centrum. The lowermost or first hypural is by far the larger of the two 
and is fan-shaped, in contrast to the parallel-sided second element. The bases of 
these hypurals are slightly constricted and produced into rounded heads which fit 
in weakly defined cup-shaped depressions on the supporting centrum. 

Three upper hypurals may be seen (though it is very likely that there was a fourth 
in the series representing the ' hypurale minimum ' of Monod 1967). The third 
hypural, the lowermost member of the upper series, is large and fan-shaped. It is 
divisible into a thickened upper and a thin lower portion. The fourth hypural is 
considerably narrower, but similar to its ventral neighbour in articulating with the 
second ural centrum. 

There are four uroneural elements. The first is expanded proximally and covers 
the dorso-lateral surface of the first and second preural centra. The second uroneural, 
which in the specimen figured (Text-fig. 53) is displaced posteriorly, is slightly 
smaller and lies in a groove on the posterior face of the first uroneural. The third 
lies in conformity with the second but the fourth appears to have been displaced 
posteriorly. The nature of this fourth element is in question. Its size and shape 
would suggest that it represents a urodermal, but its position within the fin-rays, 
not superficial to them, and the nature of the bone surface are features of a uroneural. 
This structure is interpreted as being a fourth uroneural. 

Between the first uroneural and the second preural neural spine there are three 
epurals. The posterior epural is the largest. The anterior epural is markedly 
curved but it is not clear how much of this curvature is due to the distortion which 
has obviously taken place. 

The base of the inner ray of the upper lobe shows a marked basal expansion as in 
many lower teleosts. A caudal scute was present both above and below the caudal 
peduncle. 

Squamation. A lateral line scale count cannot be made. The type specimen 
exhibits 38 scales to the level of the pelvic fin insertion. If the pelvic fin origin be 



ELOPIFORM FISHES 




00 

•*■ 

X 

m 
e 

o 
u 

i 



X! 

SP 
a 

« 

o 

i— I 

3 

o 



ts 



CO 

in 

6 



FOSSIL AND RECENT in 

taken as representing the mid-body level then the total lateral line count must have 
been about 80 (Woodward 1901 : fig. 2, indicates 87). The transverse count is 7/6 
at both the occiput and the dorsal fin. 

The scales have a characteristic shape and together with those of 0. levis are dis- 
tinctive among remains of the English Chalk fish fauna. Each scale is large, 
angular anteriorly but smoothly rounded posteriorly. The exposed portion is 
marked by fine tuberculations set in an apparently irregular fashion. The covered 
portion bears many smaller tuberculations and these are set in a radial pattern. 
Fine circuli are arranged concentrically around a central nucleus. The majority of 
the anterior body scales are marked by 3-8 (mode 5) radii which produce a scalloped 
margin anteriorly. Towards the tail the scales appear more rounded and do not 
bear radii. Lateral line scales are marked by a longitudinal ridge which terminates 
as a notch on the posterior margin of the scale. 



Osmeroides levis Woodward 
(Text-fig. 54) 

1895 Aulolepis typus Agassiz ; Woodward : 660, pi. 43, figs. 2, 3, 5 (errore). 

1900 Osmeroides sp. ; Woodward : 325. 

1 90 1 Osmeroides levis Woodward : 15. 

1907 Osmeroides levis Woodward; Woodward : 118, pi. 23, figs. 9-11. 

Diagnosis (emended). Osmeroides reaching a maximum length of 200 mm. 
Length of cranium not exceeding twice its maximum width at the sphenotic region. 
Parietals as long as wide. Dermal skull bones smooth, ornamented only by weakly 
developed radiating ridges. Interorbital septum partially ossified. Parasphenoid 
ending beneath mid-otic region. Quadrate/mandibular articulation beneath middle 
of orbit. Branchiostegal rays smooth, exact number not known but not less than 
15. Vertebrae, dorsal fin and scales as in the type-species. 

Holotype. B.M.N.H. P. 5681, from the Upper Cenomanian (Holaster subglobosus 
zone) of Burham, Kent, England. 

Material. The holotype and the following B.M.N.H . specimens ; 49903 
(paratype), P. 1854 (part and counterpart), P. 36204. The last specimen was prepared 
in acid by Dr C. Patterson. All specimens are from the Upper Cenomanian of 
S.E. England. 

Remarks. 0. levis is both rarer and more restricted in time than the type-species. 
Although the specimens are not accompanied by horizon data the matrix in which 
they are embedded is easily identifiable as having the lithology of the Holaster 
subglobosus zone, and is therefore of Upper Cenomanian age. 

Other than the cranium very little is known of the anatomy. However, sufficient 
evidence remains to indicate that 0. levis is very closely related to the type-species. 

Woodward (1907) noted the following differences between 0. levis and the type- 
species. The head and opercular apparatus occupy one-third of the standard length 
(taken in this instance as the distance between the snout and the caudal peduncle), 
whereas the head and opercular apparatus are relatively shorter in the type-species. 



112 



ELOPIFORM FISHES 

de 



sosc 




ot s.c 



soc 



10 mm 



Fig. 54. Osmeroides levis Woodward. Neurocranium in dorsal view. Based on 
B.M.N. H. P.5681, P.1854 and 49903. 



The cranial roof is shorter in proportion to its width than in the type-species and the 
parietal bones are shorter than long. Perhaps the most easily recognizable differ- 
ence noted by Woodward concerns the ornamentation of the skull roofing bones, 
those of 0. levis being smooth except for a small patch of rugose ornamentation on 
the frontal bones above the orbit, whereas the roofing bones of the type-species are 
uniformly ornamented from the mid-orbital region posteriorly. 

In addition to the above differences several others may be mentioned. The 
dermethmoid of 0. levis is considerably shorter than that of the type-species and this 
is related to the preorbital length which is less in 0. levis. 

The parasphenoid of 0. levis is relatively narrow and terminates beneath the 
middle of the otic region of the neurocranium, contrasting with the very broad, long 
parasphenoid of 0. lewesiensis. In both species the parasphenoid bears a patch of 
very small, rounded teeth. The size of the parasphenoid has resulted in other 
differences between the two species. The parasphenoid rests against the lateral 
ethmoid in both, but in 0. lewesiensis there is a sutural connection between these 



FOSSIL AND RECENT 113 

elements, whereas in 0. levis there is an interspace. The interorbital septum is not 
completely ossified in 0. levis in contrast to the well-ossified septum in 0. lewesiensis. 
Posteriorly the myodome is open in 0. levis whereas in the type-species it is closed. 

Apart from the differences noted above, the neurocranium is very like that of the 
type-species. Woodward (1907 : 118) described much of the cranial anatomy 
other than the neurocranium, and the following notes, based upon an examination 
of B.M.N.H. P. 36240, are intended to supplement that work. 

Hyopalatine bones. The hyomandibular tapers from a broad undivided head to a 
vertically directed shaft, unlike the forwardly directed shaft of the type-species. 
The opercular process is large, rounded and bears a stout horizontal ridge upon its 
medial surface. This ridge provided a point of insertion for the levator opercularis 
muscle. Laterally the shaft of the hyomandibular bears a stout ridge behind which 
there is a well-marked groove. Two foramina open into the groove from a single 
intraosseous canal, the upper representing the point of exit of the hyoidean nerve, 
the lower the mandibular nerve. The nerves continued within the hyomandibular 
groove. As in the type-species, the anterior margin of the hyomandibular is pro- 
duced into a thin wing which is partly overlain by the metapterygoid. 

The remainder of the hyopalatine series is very similar to that of 0. lewesiensis, 
the only notable difference being that the quadrate is deeper than long in 0. levis, 
which is the converse of the type-species. 

The hyoid arch and branchiostegal rays. The interhyal was not identified in any 
specimen, although its presence is indicated by a notch upon the dorsal surface of 
the posterior ceratohyal. The posterior ceratohyal is approximately triangular, 
smooth upon the medial surface, but marked laterally by two depressions separated 
by an antero-ventrally inclined ridge. The anterior ceratohyal is elongate and 
waisted in the anterior third. It is deepest posteriorly where it contacted the 
posterior ceratohyal through a narrow band of cartilage. A prominent groove 
close to the dorsal margin of the lateral surface of the anterior ceratohyal marks 
the path of the afferent hyoidean artery. Ventrally the anterior ceratohyal shows a 
groove flanked medially by a notched ledge. Each of these notches marks the point 
of insertion of a branchiostegal ray suggesting that not less than five branchiostegals 
were attached along this margin. 

There were two hypohyals, although only the upper is known. This is cuboid in 
shape and was separated from both the ceratohyal and lower hypohyal by cartilage. 
The groove for the afferent hyoidean artery runs along the dorsal surface of the 
upper hypohyal. In the type species the artery runs through the upper hypohyal. 

The median urohyal is ' T '-shaped in cross-section. Anteriorly the horizontal 
lamina of bone is constricted before becoming expanded into two antero-laterally 
directed prongs. The median vertical flange of bone is thin, but its posterior extent 
is not known. 

Not less than 15 branchiostegal rays are present. The anterior branchiostegals 
are narrow but the posterior rays are broad and marked by a stout ridge which runs 
down the centre of their length. 

Gill arches. Very little is known of the branchial arches in this or the type-species. 
The basihyal and basibranchials were covered by two large tooth plates, the basihyal 

8 



U4 ELOPIFORM FISHES 

plate and an elongate basibranchial plate representing the compound plate asso- 
ciated with the first three basibranchials. The teeth borne by these plates are small, 
set in shallow sockets and range in shape from conical to hemispherical. On the 
basibranchial plate the teeth situated anteriorly are slightly larger than those 
posteriorly. 

The cerato- and epibranchials are long, thin, and bear a deep groove upon the 
lateral face for the branchial arteries. The medial surface of these elements is 
smooth and the associated tooth plates lay free in the dermis. The tooth plates are 
oval in shape and bear villiform teeth. Somewhat larger tooth plates may be 
identified as infrapharyngobrancial tooth plates (upper pharyngeals of Nelson 
1969a). The teeth borne by the upper pharyngeals are smaller than those on the 
basibranchial tooth plates. 

Postcranial skeleton. What little is known of the postcranial skeleton is very like 
0. lewesiensis. The squamation is similar to that of the type-species but, as may be 
expected in a smaller fish, the scales show much less ornamentation. 



Osmeroides latifrons Woodward 

(Text-figs. 55-59) 

1907 Osmeroides latifrons Woodward : 119, pi. 24, figs. 1, 2, 3. 

Diagnosis (emended). Species known only from the cranium and anterior part 
of trunk. Length of cranial roof less than twice the maximum width, at the 
sphenotic level. Parietals broader than long. Frontals becoming narrow above 
lateral ethmoids. Dermal cranial bones smooth, marked only by fine radiating 
growth lines. Interorbital septum absent or membranous. Parasphenoid reaching 
rear of neurocranium. Quadrate/mandibular articulation beneath lateral ethmoid. 
Branchiostegal rays less than 15 in number. Scales large, exposed area marked 
by a fine reticulation of rugae. Anterior field of scales without radii. 

Holotype. B.M.N.H. P. 10465, exhibiting skull roof, from the Upper Cenomanian 
of Kent, S.E. England. 

Material. In addition to the type, B.M.N.H. specimens P.5679, P.9699, P.10466 
and P. 1 1 190 were examined. The last-mentioned specimen was prepared in acetic 
acid. All specimens are from the Upper Cenomanian of Kent, S.E. England. 

This species is known only by imperfect cranial remains and the anterior region 
of the trunk. The account of the cranium is based mainly on B.M.N.H. P.11190. 

Description. Osmeroides latifrons differs considerably from other species of 
Osmeroides. Many of the observable differences are features in which 0. latifrons 
resembles the Albulidae and Pterothrissidae. However, a basic similarity in neuro- 
cranial structure and the retention of several primitive features justify the inclusion 
of this form within the genus Osmeroides. The description given below is designed 
to illustrate the differences between 0. latifrons and the type-species, 0. lewesiensis. 

Neurocranium. The cranial bones are unornamented, except for a few lines of 
growth on the frontal, infraorbitals and operculum. Comparing the neurocranial 



FOSSIL AND RECENT 



"5 




asp 



soc 



Fig. 55. Osmeroides latifrons Woodward. Neurocranium in dorsal view. 
Based on B.M.N.H. P.10465. 



views (Text-figs. 55 and 46, 56 and 47, 57 and 48), the differences in proportion are 
obvious. The significant features of the neurocranial roof are : the narrow, slightly 
elongated snout, the broad otic region, the irregular shape of the parietals and the 
long dermethmoid. The dermethmoid shows a prominent median ridge which is 
continued posteriorly on the frontals. At the anterior end of the dermethmoid there 
are inclined struts of bone connecting the horizontal and vertical portions of this 
element ; in this the dermethmoid resembles that of albulid and pterothrissid fishes. 
The frontal of 0. lewesiensis is flat throughout but in 0. latifrons the centre of ossi- 
fication is raised and there is a shallow trough developed anteriorly, features seen 
in the albulids and, to a lesser degree, in the pterothrissids. 

The supraoccipital of 0. latifrons is relatively small and there is a small, deep 
depression on either side of the weakly developed crest. The sub-epiotic fossa is 



n6 



ELOPIFORM FISHES 




pto 



f m 



f uv 



Fig. 56. Osmeroides latifrons Woodward. Neurocranium in posterior view, 
This and Text-figs. 57 and 58 from B.M.N.H. P. 11 190. 



very large and the ridge which forms the medial margin in all Osmeroides species is 
particularly well developed here. The intercalar was not seen but the surface mark- 
ings on the exoccipital and pterotic suggest that it was of limited size (Text-figs. 
56, 57 and 58). 

The lateral surface of the neurocranium shows a large otic bulla and a deep sub- 
temporal fossa. Both of these structures are developed to a greater degree than in 
the type-species. Indeed, the subtemporal fossa is so deep that it severely restricts 
the lumen of the post-temporal fossa. Also seen upon the lateral face of the prootic 
is a prominent ridge of bone which passes antero-ventrally from the level of the pars 
jugularis. This ridge, which provided a site of origin for branchial musculature, 
is present but less pronounced in other species of Osmeroides. 

The trigemino-facialis region is basically similar to that of the type-species, but 
there are two small differences. Firstly, the profundus ciliaris of 0. latifrons leaves 
the cranial cavity by a separate foramen, whereas in 0. lewesiensis that nerve ran 
out from the cranial cavity with the trigeminal nerve. The second difference 
concerns the path taken by the orbital artery. The condition in 0. lewesiensis 
has been described (p. 99) and is quite typical of elopiforms. In 0. latifrons (as 
represented by B.M.N.H. P.11190) an unusual condition existed. The orbital 
artery pierced the lateral face of the prootic and turned antero-dorsally to pass into 
the orbit immediately beneath a small bridge of bone, without taking its usual path 
through the pars jugularis. The path of the orbital artery presumably resulted 
in the efferent hyoidean branch looping back into the pars jugularis before running 
out of the prootic with the hyomandibular ramus of the facial. Since only one 



FOSSIL AND RECENT 



117 



suitable specimen of 0. latifrons was available it is not certain if the unusual disposi- 
tion of the orbital artery is normal for this species or merely an individual variant. 

Within the orbit 0. latifrons differs from the type-species in the membranous inter- 
orbital septum and the small basisphenoid stem, although, as in 0. lewesiensis, 
there are two posteriorly directed prongs arising from the posterior margin of the 
stem. The parasphenoid of 0. latifrons is proportionally wider throughout, par- 
ticularly beneath the otic region. The parasphenoid tooth patch is as extensive 
as in the type-species but the teeth are slightly different. The parasphenoid teeth 
of 0. latifrons are of irregular size and vary in shape from small, villiform teeth to 
larger hemispherical teeth. The teeth become larger posteriorly and those in the 
centre are larger than the marginal teeth. 

Hyopalatine bones. The quadrate/mandibular articulation lies beneath the lateral 
ethmoid, further forward than in the type-species. A distinctive feature of the 
hyomandibular is the thin bone forming the head ; otherwise there is agreement in 
form with 0. lewesiensis. The metapterygoid is imperfectly preserved dorsally, 
but it probably reached the antero-dorsal point of the hyomandibular and resulted 
in the presence of a hyomandibular-metapterygoid foramen as in albulids and 
pterothrissids. The lateral surface of the metapterygoid is marked by a prominent 
ridge, present in other species of Osmeroides, which provided an anchorage for a 
superficial division of the levator arcus palatini musculature. The quadrate, 
although of normal shape, is positioned so that the posterior margin lies almost 
horizontally. There is a considerable gap between the ventral end of the hyomandi- 
bular and the quadrate. This gap was presumably bridged by the symplectic. 
The symplectic was not seen but the impressions left upon the quadrate suggest 
that it was somewhat flattened, as in Albula. The metapterygoid contributes to the 
symplectic support. The ectopterygoid is almost as wide as the endopterygoid but 



fahm 



20mm 




boc 



Fig. 57. Osmeroides latifrons Woodward. Neurocranium in left lateral view. 



n8 



ELOPIFORM FISHES 



fica 




2Hhm 



fahm 



boc 



fuv 



Fig. 58. Osmeroides latifrons Woodward. Neurocranium in ventral view, 



FOSSIL AND RECENT 119 

unlike the latter is vertical. The dorsal margin of the lateral face of the ectoptery- 
goid is produced as a small crest. The teeth borne by the ectopterygoid and endo- 
pterygoid form a single large patch. Unlike those in 0. lewesiensis, the teeth in this 
patch vary considerably in size and shape. Those situated medially resemble the 
larger parasphenoid teeth. Those found laterally are similar to the villiform teeth 
seen in the type-species. The anterior end of the hyopalatine series is not known. 

Mandible. The mandible is short, turned inwards markedly at the symphysis 
and with a posteriorly situated coronoid process. Ventrally the dentary is marked 
by a deep groove which runs parallel to the inflected margin, and carried the sensory 
canal. Posteriorly the mandibular sensory canal ran through the articular. The 
oral surface of the dentary bears a broad band of teeth. Each tooth is pointed and 
curves inwards distally like those of 0. levis but unlike the villiform teeth of 0. 
lewesiensis. The articulatory facet is formed mainly by the endosteal articular 
which is relatively larger than that element in the type-species. 

Circumorbital series. The anterior half of the first and the third, fourth and fifth 
infraorbitals are the only elements of this series known. Unlike those of the type- 
species, all bones are thin and marked only by growth lines. The first infraorbital 
shows a rounded anterior margin and is without the small dorsal prominence seen in 
0. lewesiensis. The surface of the first infraorbital is marked by concentric growth 
lines and in the dorsal half there is a thin ledge which protected the sensory canal. 
The third, fourth and fifth infraorbitals form the posterior border of the orbit. These 
elements are narrower than those of 0. lewesiensis, leaving more of the cheek region 
exposed. 

Hyoid arch and branchiostegal rays. Both anterior and posterior ceratohyals are 
short and deep, contrasting with those of the type-species. The posterior ceratohyal 
carried five branchiostegals and at least six are attached to the anterior ceratohyal. 
All branchiostegals are smooth but otherwise they resemble those of 0. lewesiensis. 

Cephalic sensory canal system. The sensory canal system is more open than in 
0. lewesiensis. The supraorbital canal ran within the frontal from the level of the 
autosphenotic to open onto the surface of the frontal at the level of the lateral 
ethmoid. Anterior to this the canal must have lain in the frontal trough. The 
infraorbital sensory canal of 0. latifrons is relatively larger than that of 0. lewesiensis 
and, unlike that of the latter species, is situated at the orbital margin. Within the 
third infraorbital the canal opens to the surface by large pores, separated from one 
another by narrow bony struts. On the first infraorbital the canal must have been 
open ventrally, the only protection being afforded by a narrow ledge of bone. The 
mandibular sensory canal was similarly open, lying as it did in a large groove. The 
only point at which the canal passes through bone is at the extreme postero-ventral 
angle of the articular. The preopercular sensory canal ran within an intramural 
canal in the dorsal half of the preoperculum. At the level of the fourth infraorbital 
the canal emerged to the surface and continued its path beneath a narrow shelf. 

Opercular series. The preoperculum is large and reaches well forward. The dorsal 
end of this element is narrow and truncated. Anteriorly the bone is rounded. The 
operculum differs in shape from that of the type-species. The anterior margin is 
almost straight while the posterior border curves backwards and downwards to near 



I2() 



KLOPIFORM FISHES 




X 

ffl 



o 
a, 



o 
o 



■5 ° 

> On 

^H M 

a! m 

V-i M 

OJ TO 

— •* 

6 2 
•3* 



o 



On 
On 

On 

>o 

vO 

o 



£ & 






o 



On 



o 

l-( 



FOSSIL AND RECENT 121 

the ventral end of the bone, where it curves sharply forwards resulting in an oblique 
ventral border. The suboperculum and interoperculum have the same form as in 
0. lewesiensis. 

Squamation. Only a few of the anterior scales are known. The scales are thinner 
than in the type-species although of the same relative size. Each scale is nearly 
circular. The anterior field is marked by random fine granulations. In the dorsal 
and ventral fields the granulations become defined as circuli while in the posterior 
field they form a network of rugae and thus differ from those in 0. lewesiensis 
which are set in radiating fashion. The scales of 0. latifrons are without anterior 
radii and the anterior margin is complete ; a few of the anterior body scales of the 
type-species are also without radii. The lateral line scales may be identified by a 
small notch in the posterior margin, as in 0. lewesiensis. 

Affinities of Osmeroides latifrons. Osmeroides latifrons is an advanced species 
of the genus. Many features of 0. latifrons are interpreted as the initiation of trends 
which are characteristic of albulids and pterothrissids. For instance, the cephalic 
sensory canals of 0. latifrons are considerably more open than those of 0. lewesiensis 
or 0. levis, but are never cavernous as in the Albulidae and Pterothrissidae. Inter- 
mediate conditions of other characters are seen in the snout, hyopalatine series and 
the dentition, and are discussed in relation to the evolution of the Albuloidei (pp. 
202-210). Despite certain resemblances to the Albulidae and Pterothrissidae, 0. 
latifrons retains primitive characters of the Osmeroididae such as the terminal 
mouth, the incorporation of the first centrum into the neurocranium, the unexpanded 
ectopterygoid process, the presence of many ectopterygoid teeth and the unspecialized 
dentition, particularly that associated with the gill arches. In all, the known 
osteological features of 0. latifrons show a closer resemblance to other species of 
Osmeroides than to any other albuloid genus. Those points in which 0. latifrons 
differs from 0. lewesiensis and 0. levis concern degree of development rather than ab- 
solute difference and it is therefore difficult to justify exclusion of 0. latifrons from 
the genus. 0. latifrons is known only by incomplete cranial remains and until more 
complete specimens are available it is retained in Osmeroides. Osmeroides latifrons 
is most nearly related to 0. levis. With this species it shares such features as smooth 
cranial bones, a very shallow otic region of the neurocranium with large otic bullae, 
a membranous interorbital septum and small dentary teeth which are decidedly 
pointed. With 0. lewesiensis it shares but one character, a very broad parasphenoid. 

The cranial anatomy of 0. latifrons suggests a morphotype for the Albulidae and 
Pterothrissidae. Evolution of the latter fishes from 0. latifrons would primarily 
involve loss of the medial ridge of the sub-epiotic fossa ; deepening of the trough 
upon the skull roof ; a general enlargement of the sensory canals with a reduction 
in their bony protection ; reduction or, in the case of the Albulidae, loss of ectoptery- 
goid teeth ; expansion of the ectopterygoid process to contact the infraorbitals ; 
development of an inferior mouth ; development of an inflected dentary margin 
with consequent approximation of the jaw rami in the ventral mid-line and reduction 
or loss of the gular plate. In three features 0. latifrons resembles the Albulidae 
rather than the Pterothrissidae ; the position of the quadrate/mandibular articula- 
tion, the broad parasphenoid and the irregular shape of the parietals. The first 



122 ELOPIFORM FISHES 

two of these similarities would not exclude 0. latifrons from the ancestry of the 
Pterothrissidae since the position of the lower jaw articulation varies within each 
albuloid family and a broad parasphenoid is a basic albuloid character, also found 
in the primitive Osmeroides lewesiensis. The shape of the parietals, however, is a 
character of the Albulidae not found in the Pterothrissidae. This similarity is noted 
here, but no conclusions are drawn, as the phylogenetic implications of parietal shape 
are not understood. 



Genus DINELOPS Woodward 
1907 Dinelops Woodward : 121. 
Diagnosis. See Woodward (1907 : 121). 
Type and only species. Dinelops ornatus Woodward. 

Dinelops ornatus Woodward 
1907 Dinelops ornatus Woodward : 121 , pi. 24, figs. 4-6. 
Diagnosis. See Woodward (op. cit.). 

Holotype. B.M.N.H. 39432, an incomplete cranium from the Upper Cenomanian 
of Kent. 

Material. Only three specimens are known, the holotype and the paratypes 
B.M.N.H. 49091 and P.1812. The specimens consist of incomplete cranial remains 
and parts of the trunk from the Lower Chalk of S.E. England. 

Remarks. The affinities of this genus are problematical. Woodward (1907) 
stated that Dinelops resembled Osmeroides, a view which, with reservation, is upheld 
here. 

The skull roof is relatively broad above the cranial vault but tapers sharply 
immediately anterior to the level of the autosphenotic spine, as in Osmeroides 
latifrons. The roofing bones are ornamented with coarse rugae which, unlike any 
species of Osmeroides, extend on to the dermethmoid. Above the cranial vault 
there is a median frontal depression but it is weakly defined. A medial branch of 
the supraorbital sensory canal opens on to the surface of the frontal above the hind 
end of the orbit and continues forward in a shallow trough, a condition similar to that 
of Albula. 

The supraorbital, as in Osmeroides, is large and ornamented posteriorly. 

The orbit is relatively large, larger than in any species of Osmeroides, and sur- 
rounded by a series of six canal-bearing infraorbitals. The sixth member, the 
dermosphenotic, is known by small fragments. The bones are thin, as in Osmeroides 
latifrons, and devoid of ornamentation. The posterior infraorbitals are wide, as in 
Elops, but the fourth is distinctively very shallow. 

Of the upper jaw only the maxilla is known. It is totally unlike that of Osmeroides. 
The maxilla is straight and extends back to beneath the centre of the posterior infra- 
orbitals. Anteriorly the maxilla is narrow and bears a prominent palatine cup 



FOSSIL AND RECENT 123 

before ending in a slightly expanded maxillary head. Woodward (1907) suspected 
the presence of only one supramaxilla, but the specimens are too imperfect to confirm 
this suspicion. The teeth are unknown but a series of moderately large alveoli 
set in a single row indicate a dentition unlike that of Osmeroides. The alveoli are 
only weakly delimited from one another and in some parts of the bone the teeth appear 
to have been set in a continuous groove. 

The quadrate/mandibular articulation occurs well behind the eye, beneath the 
occiput. The mandible is deep throughout and a well-developed coronoid process 
is absent. B.M.N.H. 49091 is the only specimen showing the dentition. At the 
anterior end of the jaw the teeth are arranged in a band about four teeth wide. 
Behind this level there is a single row of prominent alveoli in which were set somewhat 
larger teeth. Lateral to this row of prominent teeth there are much smaller teeth, 
while towards the anterior end of this series small alveoli indicate that there may have 
been a medial row of very small teeth. The teeth are pointed. The absence of jaws 
in the type specimen makes positive identification of this jaw (B.M.N.H. 49091) 
difficult (Woodward 1907 only tentatively assigned this specimen to Dinelops). 
However, the ornamentation upon the mandible is very similar to that of the 
roofing bones. 

The post cranial skeleton is even more poorly known. There are approximately 
14 rays in the pectoral fin which was held low down on the body. The dorsal fin 
was situated above the pelvic fins and relatively far back. 

The scales are large, almost circular and marked only by fine circuli. A few 
scales exhibit fine granulations in the posterior field. 

The characters of the skull roof and supraorbital agree with those of the Osmeroi- 
didae. The jaws and dentition are, however, decidedly different and more advanced 
than Osmeroides. Failing knowledge of the neurocranium, Dinelops is tentatively 
retained in association with Osmeroides. 



Family PTEROTHRISSIDAE Gill, 1893 

Diagnosis. Albuloid fishes in which the snout is elongated ; mouth inferior. 
Ethmoid commissure incomplete, running through premaxillae. Supraorbital 
sensory canal not extending on to parietal. Occipital condyle formed entirely by 
basioccipital. Interorbital septum membranous. Lateral ethmoid not sutured 
to the parasphenoid, which is narrow. Parasphenoid and endopterygoid with 
conical teeth opposed by a similar dentition upon the basihyal and basibranchial 
tooth plates. Vomer edentulous. Premaxilla, maxilla and dentary with small, 
needle-like teeth. Posterior infraorbitals narrow, incompletely covering the 
cheek. Supraorbital small. Hyomandibular-metapterygoid foramen developed. 
Premaxilla tightly bound to the mesethmoid ; maxilla moving independently ; a 
single ovoid supramaxilla. Mandible with prominent coronoid process. Branchio- 
stegals less than 10 in number. Gular plate absent. Dorsal fin elongated, anal fin 
short based. Caudal skeleton with two or three uroneurals. Inner caudal rays of 



124 ELOPIFORM FISHES 

each lobe without expanded bases. Principal caudal rays preceded by numerous 
basal fulcra, many of which are unsegmented and lie free in the body musculature. 
Caudal scutes absent. 



Genus PTEROTHRISSUS Hilgendorf, 1877 

1877 Pterothrissus Hilgendorf : 127. 
1877 Bathythrissa Giinther : 443. 

Diagnosis. Pterothrissidae in which the interorbital distance is approximately 
one-quarter of the length of the neurocranium. Parasphenoid teeth extending to 
mid-orbital level. Maxilla with 6-7 teeth. Coronoid process forwardly directed 
and situated in the anterior third of the mandible. Branchiostegals six in number. 
Vertebral column with never less than 100 vertebrae. 

Type-species. Pterothrissus gissu Hilgendorf. 

Remarks. The genus is known by two species, the type-species, occurring off 
the coasts of Japan, and P. belloci Cadenat from off tropical West Africa. Both 
species are inhabitants of deeper waters. Poll (1953) records that most specimens 
of P. belloci have been found at the edge of the continental shelf. Information on 
P. gissu is imprecise but Okada (i960) states that it is found in ' moderately deep 
water '. Specimens of P. belloci were not examined but it is obvious from the 
description given by Poll (1953) that it closely resembles P. gissu. The main 
differences appear to be in the relatively greater length of the head and eye in P. 
belloci, and the lateral line scale count (85-90 in P. belloci, 99-112 in P. gissu). 

Previous osteological information on Pterothrissus is scanty. Ridewood (1904) 
briefly noted some differences between Pterothrissus gissu (termed Bathythrissa 
dorsalis in that publication) and Albula. Gosline (1961) referred to the caudal 
structure and snout, and Monod (1968) described and figured the tail of P. belloci. 



Pterothrissus gissu Hilgendorf 

(Text-figs. 60-69) 

1877 Pterothrissus gissu Hilgendorf : 127. 
1877 Bathythrissa dorsalis Giinther : 443. 

Diagnosis. Pterothrissus reaching 400 mm S.L. Length of head equal to 
25-26 per cent of the standard length. Maximum depth of the head at the occiput 
and equal to 50 per cent of the head length. Diameter of orbit 27-30 per cent of 
head length. Dorsal fin composed of 58-60 rays, anal 11-12 rays, pectoral 15 rays, 
pelvic 10 rays. 99-112 scales in lateral line. Transverse scale count at dorsal 
fin origin is 7 above and 8 below the lateral line. 

Material. Four specimens were used in the preparation of the following osteo- 
logical description. Three were donated by Prof. T. Abe of Tokyo University 
(221 mm, 236 mm and 220 mm S.L.). The fourth is a larger specimen, B.M.N.H. 
1903.5. 14.142. 



FOSSIL AND RECENT 125 

Description. Neurocranium. The neurocranium is shown in dorsal, lateral, 
ventral and posterior views in Text-figs. 60, 61, 62 and 63. It is long and shallow 
with a slightly elongated snout which is turned downwards at its tip imparting to the 
head a characteristic albuloid appearance. Cartilage forms much of the ethmoid and 
nasal capsules and occupies large interspaces between certain bones in the otic region. 

The ethmoid consists of a median vertical plate of cartilage which expands an- 
teriorly to meet the contours of the dermethmoid above and the vomer below. Along 
the dorsal surface of the ethmoid there is a thin layer of perichondral bone. Dorsally 
this mesethmoid is overlain by a large dermethmoid, an element formed by a 
horizontal plate with a prominent vertical ridge on the dorsal surface. The ridge is 
pierced by a large fenestra, while anteriorly inclined struts link the vertical ridge 
with the horizontal portion. Beneath each strut there is an articular facet receiving 
the head of the premaxilla. The facet is formed by the perichondral mesethmoid. 

The vomer is edentulous and narrow, unlike that element in albulids and os- 
meroidids. Anteriorly the head of the vomer is slightly expanded and is not clearly 
separable from the mesethmoid. The head of the vomer is produced ventrally as 
a stout median ridge which is semicircular in lateral view. 

Each lateral ethmoid is represented by thin perichondral bone which covers the 
posterior and part of the lateral wall of the nasal capsule. Laterally the lateral 
ethmoid bears a stout ridge which receives ligaments from the posterior process of the 
palatine. A single foramen in the transverse wall of the lateral ethmoid transmits 
the olfactory tract, nasal artery and the superficial ophthalmic branches of V and 
VII. The lateral ethmoid is separated from its fellow by a narrow strip of ethmoid 
cartilage. 

The frontal is narrow anteriorly where it is raised in the mid-line to form a ridge 
lying in continuity with the ridge on the dermethmoid. Above the orbit the lateral 
margin of the frontal forms a structure resembling a mudguard. Medial to this 
' mudguard ' there is a shallow trough which ends posteriorly in a large sensory pore. 
The nasal lies above the anterior end of the frontal and the posterior region of the 
dermethmoid. Each nasal is formed by an anterior and larger posterior ossicle, 
each barely larger than the contained sensory canal. Both ossicles are open dorsally 
along their entire length. The presence of two ossicles composing the nasal element 
is noted elsewhere among elopiforms (Elops, Megalops and Tarpon) and is a reflection 
of a general elopiform tendency towards fragmentation of the snout elements. 

The rectangular parietal is longer than broad and meets its partner over much 
of its length. A rebate developed along the posterior margin of the parietal contains 
the central portion of the supratemporal commissure. The parietal is perforated 
by several small pores (Text-fig. 60) and although the precise disposition of these 
pores varies from specimen to specimen there appears to be a longitudinal and a 
transverse series. Such an arrangement of pores is similar to those within the 
anterior and middle pit-lines of halecostomes, but in the latter fishes the pores are 
linked by shallow grooves, not seen in Pterothrissus. 

The dermal portion of the pterotic forms the dorso-lateral corner of the skull roof. 
The otic sensory canal ran within a shallow groove along the lateral edge of the 
pterotic. A narrow rebate along the posterior margin of the pterotic, which lies 



126 



ELOPIFORM FISHES 



ot.s.c 




stt.com 



stt.com 



Fig. 60. Pterothrissus gissu Hilgendorf. Neurocranium in dorsal view. 
This and Text-figs. 61-69 are composites from several specimens. 



FOSSIL AND RECENT 127 

adjacent to an open groove upon the supratemporal, contains the lateral portion of 
the supratemporal commissure. In lateral view (Text-fig. 61) the endochondral 
portion of the pterotic is seen to form the greater part of the dilatator fossa, the pos- 
terior half of the hyomandibular facet and the roof of the deep subtemporal fossa. 
Beneath the hyomandibular facet the pterotic is marked by an arc-shaped swelling 
which marks the path of the horizontal semicircular canal. The medial surface of 
the pterotic forms the roof, the lateral wall and part of the floor of the post-temporal 
fossa. 

In dorsal view the supraoccipital (Text-fig. 60) is seen to separate the parietals 
posteriorly. Two antero-laterally directed wings of the supraoccipital extend beneath 
the parietals and contact the autosphenotic. The anterior wings also form the 
inner wall of the post-temporal fossae. In posterior view (Text-fig. 63) the supra- 
occipital is small, semicircular and separated from the exoccipital by a large area of 
cartilage. The supraoccipital crest is very small. 

Between the supraoccipital and pterotic lies the irregularly shaped epiotic (Text- 
figs. 60, 63). A narrow band of cartilage separates the epiotic from the exoccipital. 
The posterior vertical semicircular canal runs through the epiotic and its path may be 
seen externally as a ridge running along the medial margin of the post-temporal 
fossa and forming an arc above the subepiotic fossa. The epiotic process is weakly 
developed. 

The exoccipital (Text-fig. 61) forms the posterior wall of the subtemporal fossa 
and part of the large otic bulla. Three foramina are present on the lateral face of the 
exoccipital : anteriorly there is a glossopharyngeal foramen which is preceded by a 
shallow groove, then follows a small foramen for a blood vessel, and posteriorly a 
large, ventrally directed vagus foramen. The posterior face of the exoccipital 
(Text-fig. 63) forms the ventral half of the subepiotic fossa. The exoccipital meets 
its fellow above and below the foramen magnum. The lateral and posterior faces 
of the exoccipital meet in a pronounced thickening, within which run the horizontal 
and posterior vertical semicircular canals. 

At the lower margin of the opening to the post-temporal fossa there is an area of 
cartilage between the exoccipital, pterotic and epiotic. Covering this cartilage is the 
intercalar which, as in all albuloids, is relatively small. The posterior surface of the 
intercalar is ridged, providing a surface of attachment for a tendon from the intercalar 
limb of the post-temporal. Laterally the intercalar curves antero-ventrally and in 
one specimen examined (Text-fig. 61) forms the dorsal margin of the vagus foramen. 

The basioccipital forms the occipital condyle, the posterior portion of the myodome 
roof and the walls of the saccular recess of either side. The vertebral-like expansion 
of the basioccipital shows a deep notochordal pit, which is not quite symmetrical 
since the ventral margin lies more horizontally than the dorsal. Between the bullate 
portion and the occipital expansion the lateral surface of the basioccipital is rough, 
providing an anchorage point for a strong ligament passing to the cleithrum/supra- 
cleithrum overlap. Ventrally the basioccipital is marked by a shallow groove which 
is continuous with a small groove upon the parasphenoid. 

The prootic is, as usual, a bone of complex shape. The lateral face contributes 
to the formation of the hyomandibular facet, the subtemporal fossa and the otic 



128 



ELOPIFORM FISHES 




•— u 






o 

C 

<u 



6 



FOSSIL AND RECENT 

vo fapmx 



129 




pto 



boc 



/ 
X 



exo 



ic 



Fig. 62. Pterothrissus gissu Hilgendorf. Neurocranium in ventral view. 



bulla. At two points the prootic fails to meet the basioccipital : anteriorly where 
these bones meet the parasphenoid and posteriorly where they meet the exoccipital. 
There are thus left two triangular interspaces, which even in the largest specimen 
examined remain open. 

A large foramen for the hyomandibular trunk of VII opens directly from the pars 
jugularis, so that when the neurocranium is viewed laterally (Text-fig. 61) the tri- 
geminal and facial foramina may be seen opening from the pars ganglionaris. The 
orbital artery enters the prootic at the base of a deep groove situated immediately 
above the ascending wing of the parasphenoid. The orbital face of the prootic is of 



i3« 



ELOPIFORM FISHES 



SOC 




Ot.S.C 



asp 



exo 



fm np 



5mm 



Fig. 63. Pterothrissus gissu Hilgendorf. Neurocranium in posterior view. 



limited extent ; it is pierced laterally by the large anterior opening of the pars 
jugularis and medially by the oculomotor foramen. The relationships of the rami 
of V and VII to the pars ganglionaris and pars jugularis are similar to those described 
for Osmeroides lewesiensis (p. 99). The prootic bridge is narrow and separated 
from the basioccipital by cartilage. The anterior margin of the bridge is indented 
where it forms the posterior half of the hypophysial foramen. Posteriorly the 
abducens nerve pierces the prootic bridge. 

The autosphenotic contacts the pterosphenoid medially and the prootic ventrally. 
Laterally the autosphenotic is produced as a short ' spine '. Postero-laterally the 
autosphenotic contributes to the dilatator fossa and the extreme anterior end of the 
hyomandibular facet. Much of the autosphenotic is overlain by the frontal. The 
otic branch of VII, which in most teleosts pierces the centre of the orbital face of the 
autosphenotic, here pierces the suture between the autosphenotic and pterosphenoid. 

The pterosphenoid lies posterior to the orbitosphenoid, antero-medial to the 
autosphenotic and medial to the prootic. Part of the border of the optic foramen 
is formed by the pterosphenoid. The orbitosphenoid is quite large and in the largest 
specimen examined (B.M.N.H. 1903. 5. 14. 142) is produced anteriorly as a single 
median plate which bifurcates distally and is connected to the lateral ethmoids by 
short ligaments. The surface of the orbitosphenoid is marked by a prominent groove 
anterior to the level where the olfactory tracts leave the cranial cavity. 



FOSSIL AND RECENT 



131 



The basisphenoid is ' Y '-shaped, each wing contacting the pterosphenoid and 
prootic in such a way as to exclude the latter from the optic foramen. The stem of 
the basisphenoid is expanded distally. 

Throughout most of its length the parasphenoid is narrow, expanding only at the 
level of the short ascending wings. Both anteriorly and posteriorly the parasphenoid 
is marked by a median groove. Towards the rear of the orbit the parasphenoid 
bears approximately 30 teeth. Each tooth is conical with a rounded tip. The 
efferent pseudobranchial and internal carotid arteries pierce the parasphenoid at the 
level of the ascending wings. 

Hyopalatine bones. The hyopalatine series of Pterothrissus is of a pattern charac- 
teristic of other pterothrissids and albulids but rather different from that in the 
primitive species of Osmeroides (0. lewesiensis and 0. levis), which resemble the 
Elopoidei in this respect. 

The hyopalatine series of Pterothrissus is long and shallow. The hyomandibular 
has a single broad articulatory head. The bone is thin in this region and this is 
reflected in the narrow hyomandibular facet on the neurocranium. A stout ridge 
runs the length of the hyomandibular shaft. Dorsally the posterior divisions of the 
levator arcus palatini muscle insert on this ridge, while ventrally the ridge serves 
as a site of origin for the A 2 division of the adductor mandibulae. Beneath the 
hyomandibular head the posterior margin is expanded to form a rounded opercular 
process. Upon the medial face a ridge is seen to grade into the opercular process ; 
to this ridge are attached the deeper fibres of the levator opercularis. The hyo- 
mandibular trunk of VII enters the hyomandibular on the medial surface near to the 
anterior end of the head. Within the bone the nerve splits to two subequal branches, 
both of which emerge on the posterior edge of the shaft. The upper, smaller, oper- 
cular branch leaves the bone immediately beneath the opercular process while the 



enp 



mpt 



hm 




Ihm 



Fig. 64. Pterothrissus gissu Hilgendorf. Hyopalatine bones of left side in lateral view. 

Cartilage hatched. 



132 



ELOPIFORM FISHES 

ynhm 



apal 6C . P 



dpa 




eart 



Fig. 65. Pterothrissus gissu Hilgendorf. Hyopalatine bones, opercular series (stippled) 
and mandible of right side in medial view. Cartilage hatched. 



lower, mandibular branch opens laterally into a shallow groove mid-way along the 
length of the shaft. A thin anterior wing is developed in the dorsal half of the hyo- 
mandibular. There is no overlap of the hyomandibular by the metapterygoid. 
Instead, there is a large foramen between these bones, the hyomandibular-metaptery- 
goid foramen, through which passes a deep division of the levator arcus palatini to 
insert upon the medial face of the metapterygoid. 

The quadrate is basically fan-shaped with the postero-ventral angle produced as 
a short spine. The ' posterior ' quadrate margin lies horizontally as in most albu- 
loids. 

The endopterygoid is somewhat longer than broad. A patch of approximately 
40 teeth, similar in shape to those on the parasphenoid, is borne by the convex 
palatal surface. The ectopterygoid of Pterothrissus, unlike that of Osmeroides, is 
narrow and only meets the endopterygoid over a short distance (Text-fig. 64) . The 
ectopterygoid bears a few tiny teeth which are difficult to see in ' unprepared ' 
specimens. Dorsally the ectopterygoid process is relatively small and is directed 
towards the infraorbital bones. A short ligament connects the process with the 
infraorbital series at the junction of the first and second infraorbitals. 

There is no ossified autopalatine. The palatine cartilage articulates with the 
mesethmoid by a simple rounded head, the anterior process of the palatine. Beneath 
the lateral ethmoid there is another dorsally directed process, the posterior process 
of the palatine. The dermopalatine is represented by a small, toothless, oval 
ossification attached to the ventral surface of the palatine cartilage. 

Dermal upper jaw. The premaxilla is stout with an inwardly turned head bearing 
a single articulatory facet. A tough ligament runs from the external face of the 



FOSSIL AND RECENT 



133 



premaxilla head to the dorsal margin of the dermethmoid ridge. This ligament, the 
tight fit of the articulatory facet and the presence of tough connective tissue combine 
to make the premaxilla practically immovable. The lateral surface of the premaxilla 
is perforated by three or four pores which lead to a sensory canal running through 
the bone. Posteriorly the alveolar process of the premaxilla is tapered and received 
in a small notch in the maxilla. 

The maxilla is narrow and stout anteriorly where it curves inwards to articulate 
with the ethmoid via a biconcave cartilaginous pad. Posteriorly the maxilla is 
thin and ovoid in lateral view, save for a small excavation into which a single oval 
supramaxilla fits. 

The majority of the upper jaw teeth are found on the premaxilla. The oval 
margin of the premaxilla is clothed with many small needle-like teeth which point 
towards the buccal cavity. The maxilla has a feeble dentition, there being only five 
or six pointed teeth set in a single row. 

Mandible. The mandible is short and deep with a prominent coronoid process 
situated anteriorly. Ventrally the margins of the dentary and articular are in- 
flected and form a broad shelf. From the lateral face of this shelf a thin flange of 
bone projects and this forms a partial protection for the mandibular sensory canal. 

The dentary is narrow anteriorly and is markedly turned inwards at the symphysis. 
In shape, size and disposition the numerous dentary teeth resemble those on the 
premaxilla. 

The articular is slightly larger than the dentary. Posteriorly it forms the outer 
aspect of the articular facet. The endochondral articular (endosteal articular of 
Ridewood 1904) is distinguishable by the spongy nature of the bone but it appears 



dsp 



hm 



m ros 




sop 



pmx 



10mm 



den 



art lo 3 



pop 



Fig. 66. Pterothrissus gissu Hilgendorf. Cranium in left lateral view. 



134 ELOPIFORM FISHES 

to be fused with the articular. The Meckelian cartilage rests on a ledge on the medial 
face of the mandible. The sesamoid articular is partially embedded within the 
cartilage. 

Circumorbital series. The circumorbital series is composed of six infraorbitals, 
three rostral ossicles, an antorbital and a supraorbital. 

The orbital margin of the six infraorbitals is rolled laterally so forming a ventrally 
incomplete tube round the sensory canal. The first infraorbital is large with a 
rounded anterior margin. Much of the maxilla is covered by this element. Pos- 
teriorly the first infraorbital partially overlies the second of the series but the 
connection between these bones is flimsy and they separate easily. As is usual in 
elopiforms the third infraorbital is the largest of the series. This element is angled 
and the posterior margin exhibits a shallow indentation. The fourth and fifth 
infraorbitals form a narrow inclined strut . The sixth infraorbital, or dermosphenotic, 
has no direct contact with any other bone but lies free in the skin. In all, the 
posterior infraorbitals are quite distinct from those of Albula, as may be seen by 
comparing Text-figs. 66 and 80. 

The antorbital is a narrow, elongate ossification which together with a similarly 
shaped supraorbital forms a flexible strut serving to help expand and contract the 
supraorbital nasal diverticulum. The importance of this strut in relation to the 
functioning of the diverticulum has been noted by Derschied (1924) and Gosline 
(1961). The three small rostral ossicles lie free from one another and from any 
neighbouring bone. The margins of each ossicle are rolled over to contain the sensory 
canal. 

Opercular series. The shape of the operculum, suboperculum, preoperculum and 
interoperculum may be seen in Text-figs. 65 and 66. The suboperculum is notable 
in having a particularly long dorsally directed anterior wing which almost reaches 
to the level of the opercular process of the hyomandibular. The interoperculum, 
which resembles a branchiostegal ray, is joined anteriorly to the rear of the mandible 
and posteriorly to the suboperculum by ligaments. The preopercular sensory canal 
runs within a short bony tube dorsally and is protected ventrally by a large flange of 
bone. The middle portion of this canal lies free in the skin. 

Cephalic sensory canal system. The sensory canals of the head are more open than 
in any other albuloid. For the most part the canals lie within troughs covered by 
taut skin. 

The supraorbital canal passes through the frontal for a very short distance, much 
of its length being accommodated within the frontal trough. Anteriorly the supra- 
orbital canal continues within the gutter-like nasal and ends blindly at the tip of 
the dermethmoid. The supraorbital canal connects with its partner of the opposite 
side through a large fenestra in the dermethmoid ridge. In two of three specimens 
examined in this respect a single neuromast was observed in the skin at the base of 
this fenestra (a feature previously noted by Gosline 1961) and this neuromast, like 
others of the supraorbital canal, is supplied by the superficial ophthalmic nerve. In 
the third specimen examined no neuromast could be seen. 

The otic sensory canal lies within a groove upon the pterotic. Anteriorly the otic 
canal joins the supraorbital and infraorbital canals in the skin immediately above the 



FOSSIL AND RECENT 135 

dermosphenotic. Posteriorly the otic canal joins the preopercular canal and the 
supratemporal commissure at the postero-lateral angle of the pterotic. The lateral 
line passes back through the supratemporal to open on the undersurface of that bone. 

The infraorbital canal runs within the large infraorbital trough. Anteriorly the 
canal splits into two branches : dorsally there is a short antorbital branch which 
runs over the antorbital and ends blindly at the dorsal end of that element ; anteriorly 
an ' ethmoid ' branch runs through the rostral ossicles before passing through the 
premaxilla to end blindly within that bone. The triradiate union of the ' ethmoid ' 
branch, antorbital branch and the main infraorbital canal lies in the skin just anterior 
to the first infraorbital. At one point, above the gap between the middle and 
posterior rostral ossicles, the ' ethmoid ' canal and supraorbital canals are only 
separated by membrane. 

The course of the preopercular sensory canal has been described above. Ventrally 
this canal continues within the deep trough upon the dentary and articular as the 
mandibular sensory canal. 

Hyoid arch and gill arches. The ceratohyal is separated by a band of cartilage 
into a semicircular posterior element and a longer, waisted anterior ceratohyal. 
Two hypohyals are represented by perichondral ossifications around a single large 
cartilage. The smaller dorsal hypohyal articulates with the posterior face of the 
basihyal and the ventral hypohyal receives a ligament from the urohyal. The 
afferent hyoidean artery pierces the medial surface of the ventral hypohyal and 
passes into a groove on the lateral surface of the anterior and part of the posterior 
ceratohyal. Six branchiostegal rays are attached to the medial surface of the 
ceratohyal elements. The uppermost (posterior) ray is the only member of the series 
associated with the posterior ceratohyal. 

The basibranchium is formed by a partially ossified basihyal and three separately 
ossified basibranchials. The fourth basibranchial is entirely cartilaginous. The 
basihyal tooth plate is large and bears approximately 15 pointed teeth posteriorly. 
A large tooth plate overlies the first three basibranchials and like the fourth basi- 
branchial tooth plate bears teeth similar to those upon the basihyal tooth plate. 
The teeth associated with the basibranchium become smaller posteriorly. 

The first two hypobranchials articulate with the posterior surface of their respective 
basibranchials. The reduced third hypobranchial articulates with the centre of its 
basibranchial. Of the five ceratobranchials the first three articulate with the distal 
ends of the hypobranchials while the fourth and fifth are connected to the fourth 
basibranchial by ligaments. 

The dorsal arch elements are represented by epibranchials, infra- and supra- 
pharyngobranchials. There are four epibranchials associated with the first four 
ceratobranchials. The first epibranchial is simple, showing a double articulatory 
head. The anteriorly directed head receives the posterior end of the small rectangu- 
lar first infrapharyngobranchial. The posterior head of the epibranchial is directed 
dorso-medially and articulates with a cartilaginous first suprapharyngobranchial. 
The second epibranchial, like the third, shows a marked bifid head with the anteriorly 
directed limb being larger than the dorso-medially directed process. The anterior 
head articulates with the posterior margin, or in the case of the third epibranchial 



136 ELOPIFORM FISHES 

with the lateral margin, of their respective infrapharyngobranchials. The anterior 
ends of the second and third infrapharyngobranchials have two heads, one directed 
antero-medially, the other directed antero-laterally and joined by cartilage to the 
epibranchial of the preceding arch. 

The fourth epibranchial is unusual in showing three dorsal processes. Basically 
this element has two heads, like the anterior epibranchials, but the posterior projec- 
tion has become subdivided. The significance of this subdivision is not clear. The 
fourth infrapharyngobranchial has the same relationship to its epibranchial support 
as those anteriorly, but is represented by a simple semi-ovoid cartilage having no 
connection with the antecedent epibranchial. 

Upper pharyngeal tooth plates are associated with the second, third and fourth 
infrapharyngobranchials. The third infrapharyngobranchial tooth plate is triangular 
while the fourth is divided into two sections, one associated more closely with the 
epibranchial, the other more closely associated with the infrapharyngobranchial. 

In addition to the basibranchial and upper pharyngeal dentition there are tooth 
plates associated with the oral surface of all cerato- and epibranchials and the first 
two hypobranchials. The tooth plates associated with these elements appear to be 
aligned in three rows, anterior, middle and posterior. The plates of the anterior 
row of the first arch are modified into gill-rakers which consist of a bulbous, toothed 
head supported on a slender stalk. The tooth plates of the posterior row on the 
fifth ceratobranchial are enlarged to form the lower pharyngeal tooth plate. Each 
tooth plate bears several pointed teeth. 

Pectoral girdle and fin. The main body of the post-temporal is represented by a 
flat plate of bone, pierced laterally by the lateral line. Two processes arise from the 
ventral surface of the main body : projecting anteriorly there is a spinous epiotic 
or dorsal limb, while ventrally there is a thicker intercalar or ventral limb. 

The supracleithrum is elongate, rounded at both ends, and shows a slight indenta- 
tion in the posterior margin. The lateral line passes obliquely through the upper 
half of the bone. The cleithrum is large. Above the fin insertion the bone is ex- 
panded to provide support for the large axillary scale. Anteriorly the cleithrum 
is tapered and is joined to its partner by a strip of tough connective tissue. Through- 
out its length the anterior margin of the cleithrum is turned medially providing a 
broad area of insertion for the sternohyoideus musculature. 

The coracoid is aliform and, unlike that of elopoids, does not extend to the tip 
of the cleithrum. Between the anterior and posterior connections of the cleithrum 
and coracoid there is a slit-like fenestra. The scapula is circular save for a deep 
notch anteriorly which represents an incomplete scapular foramen. The meso- 
coracoid is stout, particularly ventrally where it contacts the scapula and coracoid 
through a large area of cartilage. As in most albuloids the anterior margin of the 
mesocoracoid arch has contact with the cleithrum over much of its length. 

All but the outermost pectoral fin-ray are supported by the four proximal radials, 
which although of basic teleostean form, articulate with cartilage and are not, as is 
usual, directly related to the scapula and coracoid. The outermost fin-ray arti- 
culates both directly and indirectly with the scapula. Between the halves of the 
outermost ray there is a small ovoid element consisting of an ossified core surrounded 



FOSSIL AND RECENT 



137 




r p 1-4 



cor 



5mm 



Fig. 67. Pterothrissus gissu Hilgendorf. Pectoral girdle of right side in medial view. 



entirely by cartilage. In shape and disposition this radial resembles members of 
the cartilaginous distal radial series. This partially ossified distal radial, together 
with the base of the outermost fin-ray, fits into the depression upon the scapula. 

Of the 16 pectoral fin-rays only the first is unbranched. Little may be said of the 
length of the rays since the tips of the fin-rays were missing in all specimens examined. 
To judge by the width, the second is the longest pectoral ray. A small pectoral 
splint bone is associated with the base of the outermost ray. The proximal end of 
the splint is turned slightly upwards and markedly inwards to contact the small 
distal radial mentioned above. It is therefore unlike the pelvic splint bone in which 
the proximal end lies free in the body musculature. 



138 



ELOPIFORM FISHES 

ep1-3 



nsp pu 




h 3-6 



-u1-2 



h1-2 



Fig. 68. Pterothrissus gissu Hilgendorf. Caudal skeleton in left lateral view. 
Arrows indicate upper and lower principal caudal fin-rays. 



Pelvic girdle and fin. The pelvic fin originates mid- way between the snout and the 
caudal peduncle. The fin is supported by a triangular pelvic bone. A narrow band 
of cartilage forms a cap along the posterior margin of the pelvic bone. Only the 
innermost pelvic radial is ossified. There are io pelvic fin-rays, the outermost 
unbranched and the third apparently the longest. A large pelvic splint is associated 
with the base of the upper half of the outermost (the first) fin-ray. 

Vertebral column. Of the 107 vertebrae 35 are caudal. Each centrum is amphi- 
coelous and pierced for the passage of the constricted notochord. Most centra 
throughout the column are marginally deeper than long but the first sixteen and the 
posterior three or four are almost twice as deep as long. The lateral surface of each 
centrum is marked by several longitudinal ridges. 

The neural arches are autogenous. Those upon the anterior 45 centra are ex- 
panded distally and bear narrow neural spines. Throughout the abdominal region 
the spines are represented by separate lateral halves. Short, stout, autogenous 
parapophyses are found throughout the abdominal region. Those posteriorly are 
somewhat longer than those anteriorly. The first complete haemal arch and spine 
occurs on the seventy-sixth centrum. 

The pleural ribs are short and slender and partially encircle the abdominal cavity. 
Epineural intermuscular bones are associated with the first 67 neural arches. Those 
associated with the first 49 are attached to the base of the neural arch, but thereafter 



FOSSIL AND RECENT 139 

they are inserted nearer to the distal end of the arch. Short epipleural intermusculars 
are associated with the first 73 parapophyses. Anteriorly there are 11 sigmoid 
supraneurals. The anterior members of the series are larger than those posteriorly. 

Median fins. The dorsal fin is elongate, equal in length to half the standard length 
and composed of 58-60 fin-rays. The first four fin-rays are unsegmented and un- 
branched. The fifth is segmented and the sixth and those succeeding it are branched 
at least twice. The fin-rays are supported by 56-58 pterygiophores. Most of the 
pterygiophores are composed of separately ossified proximal, middle and distal 
radials but the first two, which between them support four fin-rays, are formed by a 
large compound radial and a separately ossified distal radial. 

The anal fin originates beneath the seventy-fifth vertebra and is composed of 
12 fin-rays of which the first three are unbranched and unsegmented, the next two 
are segmented but unbranched while the rest are both segmented and branched. 

The caudal fin is deeply forked and the margin of each lobe is rounded. Eight 
centra, two ural and six preural, are involved in the support of caudal fin-rays. The 
ural centra are as long as deep, sharply upturned and show little or no ornamentation. 
The neural spines of the third to sixth and the haemal spines of the first to sixth 
preural centra are longer and stouter than their anterior counterparts. These 
structures also show the development of median wings basally. The neural spine 
of the second preural centrum varies ; Gosline (1961) records a complete neural 
spine but in the four specimens I examined there is only a half spine, as in the 
specimen of P. belloci illustrated by Monod (1968). The neural arch of the first 
preural centrum is represented by a thin ossified structure above which there is a 
broad plate of cartilage which lies between the halves of the first uroneural posteriorly. 

There are two narrow uroneurals. The first is long and the expanded proximal 
end partially covers the first preural centrum. The second uroneural is shorter, 
fails to overlap any centrum and extends beyond the distal tip of the first. Such a 
uroneural disposition is typical for more advanced albuloids. 

There are six hypurals. The first and second hypurals articulate with the first 
ural centrum, the third and fourth are associated with the second ural centrum, 
and the fifth and sixth lie free. The base of the third hypural is expanded and 
partially encircles the centrum. The large gap between the second and third hypurals 
is a typical feature of the pterothrissid caudal skeleton. The epurals are represented 
by three splint-like elements forming a graded series. 

As usual in ' lower ' teleosts there are 19 principal caudal fin-rays. The bases of 
the innermost principal rays are not expanded and do not overlap the hypurals to 
any significant degree (cf. elopoids). There are 14-16 basal fulcra dorsally and 7-8 
ventrally. The anterior members of both series lie free in the musculature. 

Squamation. The body is covered with cycloid scales. There are approximately 
100 in the lateral line series while the transverse count at the origin of the dorsal fin 
is 7 above and 8 below the lateral line. Precise counts are difficult because of the 
condition of the specimens examined. Large axillary scales are associated with the 
pectoral and pelvic fins. 

Each scale is generally rounded but has a straight anterior margin. The posterior 
margin is thin and frequently split. Lateral line scales, apart from a slight anterior 



140 



ELOPIFORM FISHES 





T3 



a 



d 

W 



o 
d 

be 




So 






H-l 

ft 



FOSSIL AND RECENT 141 

prominence, differ little from the other body scales. The position of the nucleus of 
the scale varies slightly according to the position of the scale on the body. Anterior 
scales show an anteriorly situated nucleus. The converse is true of those scales 
situated posteriorly. Five to eight short radii are found within the anterior field 
of the scale but they do not reach the nucleus. Between the radii the anterior edge 
of the scale is slightly scalloped. Radii are less numerous on lateral line scales and 
the margins are generally smoother. The dorsal, ventral and anterior fields show 
well-marked circuli. In the anterior field the circuli tend to break up to form 
numerous backwardly pointing tubercles. All scales have enclosed bone cells 
which are generally more numerous towards the edges of the anterior, dorsal and 
ventral fields. 



Genus ISTIEUS Agassiz, 1844 

1844 Istieus Agassiz : 5, pt. 2, p. 91. 

Diagnosis. Pterothrissid fish in which the interorbital distance is equal to one- 
third of the length of the neurocranium. Parasphenoid tooth patch extending from 
beneath the lateral ethmoid to the level of the parasphenoid ascending wings. 
Maxilla bearing many teeth. Mandible with a weakly developed coronoid process 
situated in the posterior third of the jaw. Dentary teeth larger than those of the 
upper jaw. 

Type-species. Istieus grandis Agassiz. 



Istieus grandis Agassiz 
(Text-figs. 70-73) 

1833-44 Istieus grandis Agassiz : 5, pt 1, p. 13 ; pt 2, p. 92, pi. 18. 

1833-44 Istieus microcephalus Agassiz : 5, pt 1, p. 13 ; pt 2, p. 94, pi. 17. 

1833-44 Istieus macrocephalus Agassiz : 5, pt 1, p. 13 ; pt 2, p. 94, pi. 16 (lower figure) (errore). 

1858 Istieus grandis Agassiz ; Marck : 246. 

1858 Istieus microcephalus Agassiz ; Marck : 246. 

1863 Istieus macrocoelius Marck : 37, pi. 4, figs. 1-5. 

1863 Istieus mesospondylus Marck : 38, pi. 5, fig. 1. 

1873 Istieus macrocoelius Marck ; Marck : 59, pi. 2, fig. 2. 

1885 Istieus macrocoelius Marck ; Marck : 253. 

1901 Istieus grandis Agassiz ; Woodward : 67. 

1954 Istieus grandis Agassiz ; Siegfried : n, pi. 2, figs. 5, 6. 

Diagnosis. Istieus reaching 550 mm S.L. Length of head 30 per cent S.L., 
maximum depth of body approximately 22 per cent S.L. Operculum one-and-a-half 
times as deep as wide. 8-10 branchiostegal rays. Vertebrae 87-92, dorsal fin-rays 
53-57- 

Holotype. Imperfect fish, Museum National d'Histoire Naturelle, Paris, from 
the Campanian of Sendenhorst, Westphalia, Germany. 



i 4 2 ELOPIFORM FISHES 

Material. Seventeen specimens in the B.M.N.H. from the Campanian of Senden- 
horst, Westphalia, Germany. The fish are mainly preserved in impression in a 
fine-grained, grey limestone. The bone, where present, is soft and poorly preserved. 
In studying the material rubber casts were used. 

Remarks. In a great many features there is close agreement between the anatomy 
of Pterothrissus and Istieus. This similarity necessitates only a brief comparative 
description of Istieus grandis. 

Description. Neurocranium. The neurocranium, like that of Pterothrissus, is 
long and shallow. The snout is less elongated than that of Pterothrissus and does 
not exhibit the sharp anterior curvature. In dorsal view (Text-fig. 70) several 
minor differences from Pterothrissus (Text-fig. 60) are evident ; the dermethmoid 
and pterotic of Istieus are relatively smaller, the parietal is only marginally longer 
than broad, and the frontal is wider, particularly above the orbit. As in Ptero- 
thrissus, the dermethmoid bears a prominent median ridge supported anteriorly by 
inclined struts, the lateral margin of the frontal is raised into a ' mudguard ' above 
the eye and the posterior margin of the pterotic shows a narrow rebate for the 
supratemporal commissure. In Istieus the nasal is represented by a long gutter-like 
element, there being no evidence of two separate ossicles as in Pterothrissus. 

A lack of detail in the snouts of all the specimens leads one to suppose that, as in 
Pterothrissus, much of the ethmoid was cartilaginous. The lateral ethmoid of Istieus 
is the same shape as that of Pterothrissus but relatively larger. 

Little is known about the lateral wall of the otic region. B.M.N.H. P. 3885a 
shows an internal view of the right side, although little bone remains. This specimen 
shows that the bulla which contained the sacculith is large, the subtemporal fossa is 
deep and there is a deep depression posterior to the ascending wing of the para- 
spenoid. In all that may be seen the lateral wall of the braincase is similar to that 
of Pterothrissus. 

The parasphenoid, as in Pterothrissus, is straight and narrow and runs from beneath 
the lateral ethmoid to beneath the middle of the otic region. Woodward (1901) 
stated that no parasphenoid teeth could be seen, but B.M.N.H. P. 3885a clearly 
shows their presence. There are many more teeth on the parasphenoid than in 
Pterothrissus. Each tooth is conical with a rounded tip. The vomer is narrow and 
edentulous but does not show the semicircular median ridge seen in Pterothrissus. 

Hyopalatine bones. The hyopalatine series, as in all pterothrissids, is longer than 
deep and the quadrate/mandibular articulation lies beneath the posterior half of the 
orbit. 

The shape of the hyomandibular is very similar to that of Pterothrissus but the 
head is formed of thicker bone. Antero-ventrally the hyomandibular is partially 
overlapped by the metapterygoid, a feature not seen in Pterothrissus. However, 
the hyomandibular-metapterygoid foramen is developed in typical fashion. As in 
Pterothrissus the inner surface of the metapterygoid is grooved for the reception of 
the flattened symplectic. 

The endopterygoid is relatively large and bears more teeth than in Pterothrissus. 
The teeth all appear to be conical but they probably varied considerably in size as 
witnessed by the irregular size of the alveoli in B.M.N.H. 35012. The topographical 



FOSSIL AND RECENT 



M3 



pto 




soc 



Fig. 70. Istieus grandis Agassiz. Neurocranium in dorsal view. 
Composite of several B.M.N.H. specimens. 



i 4 4 ELOPIFORM FISHES 

relationships of the metapterygoid to the endopterygoid and ectopterygoid are 
similar to those in Pterothrissus. The ectopterygoid bears few teeth and these are 
considerably smaller than those on the endopterygoid or parasphenoid. The quad- 
rate differs from that of Pterothrissus in being without a posteriorly directed spine. 
The palatine was cartilaginous and covered anteriorly by a small dermopalatine. 

Dermal upper jaw. The upper jaw extends back to behind the level of the lateral 
ethmoid, in contrast to that of Pterothrissus. In shape and mode of articulation 
both the premaxilla and maxilla are similar to those in the Recent genus. However, 
the maxilla of Istieus bears more teeth and the single ovoid supramaxilla is rela- 
tively larger. The lateral surface of the premaxilla shows one or two small sensory 
pores. 

Mandible. The lower jaw of Istieus is relatively longer than that of Pterothrissus 
and results in the mouth being sub-terminal instead of decidedly inferior. Other 
differences between the lower jaw of Istieus and Pterothrissus include : the position 
of the coronoid process which is situated posteriorly in Istieus, the relatively smaller 
size of the articular, and the size of the dentary teeth. In Pterothrissus the dentary 
teeth are needle-like and of the same size as those on the premaxilla. Those of 
Istieus are conical (it is admitted that the difference between conical and needle-like 
is relative rather than absolute), larger and more robust than the premaxillary teeth 
(Text-fig. 71). 

Further features to be noted concerning the lower jaw of Istieus are the sites of 
ligament insertion. Immediately anterior to the articulatory cup the posterior 
margin of the articular is thickened for the insertion of the maxillary-mandibular 
ligament. Finally, the posterior face of the articular (that area which may be termed 
the retroarticular process, without the implication that a separate ossification is 
present) is pitted to receive the interopercular-mandibular ligament. 

Circumorbital series. This series is poorly known. The condition of preservation 
makes it difficult to identify a supraorbital and the rostral ossicles, of which there 
were probably three, have become so displaced as to render interpretation of their 
spatial relationships impossible. The antorbital is a narrow splint-like element. 

The infraorbital series is shown in Text-fig. 71. Although the suture lines between 
the separate elements are not distinct on any specimen an overall resemblance with 
Pterothrissus is evident. Certain minor differences occur : the first infraorbital of 
Istieus grandis is relatively smaller, less rounded and does not bear a dorsal promi- 
nence ; posteriorly, at the ' elbow ' of the series, the margin is not produced in a 
process and the sensory canal appears to have lain entirely superficial to the fourth 
and fifth infraorbitals. These minor differences are overshadowed by the strong 
points of similarity such as the small dermosphenotic, which is more closely asso- 
ciated with the skull roof than with the other infraorbitals, the narrow posterior 
infraorbitals and the large rolled margin. 

Opercular series. A comparison of Text-figs. 71 and 66 shows a close similarity 
between the constituent elements in Istieus and Pterothrissus. Two minor differ- 
ences are apparent : the operculum of Istieus is smaller and the postero-ventral 
part of the preoperculum is less expanded. The anterior limb of the suboperculum 
of Istieus is smaller than in Pterothrissus. 



FOSSIL AND RECENT 



H5 




i 4 6 ELOPIFORM FISHES 

Cephalic sensory canal system. The sensory canals of the head were mainly con- 
tained within open troughs. A comparison of Istieus with Pterothrissus shows a 
similarity in basic design. On the skull roof the supraorbital canal of Istieus runs 
for a greater distance within the frontal, opening to the surface of the bone above 
the anterior half of the orbit. The otic sensory canal of Istieus is contained within 
a groove, the walls of which are more complete than in Pterothrissus. 

Hyoid arch and branchiostegal rays. The anterior ceratohyal of Istieus is relatively 
longer than in Pterothrissus and this is reflected in a greater number of branchiostegal 
rays. 8-10 branchiostegals are present in Istieus, the upper (posterior) two being 
borne by the posterior ceratohyal. The variation in the number of branchiostegals 
affects those placed anteriorly. 

Each branchiostegal is broad, curved and marked by a prominent ridge which 
expands proximally to form an articular head. 

Poster anial skeleton. The impressions left by the pectoral girdle and fin are 
imperfect but the observable details are similar to comparable structures in Ptero- 
thrissus. The supratemporal is small and placed laterally, the cleithrum shows a 
posterior expansion above the area of fin insertion and the anterior portion of the 
cleithrum is narrow. The fin is composed of not less than 12 rays. In outline the 
fin is rounded distally. A pectoral splint is associated with the base of the outermost 
ray. 

The pelvic girdle and fin are closely similar to those of Pterothrissus. The pelvic 
fin originates beneath the thirty-eighth to forty-second vertebra. 

The vertebral count varies from 87 to 92 in the few specimens from which reasonably 
accurate counts could be made. The posterior 38-40 vertebrae are caudal, giving a 
caudal/total ratio of 48 per cent, a figure considerably higher than in Pterothrissus 
gissu in which this ratio is 32 per cent (Istieus macrocephalus has a caudal/total ratio 
of 40 per cent). Even though there are relatively more caudal vertebrae in /. 
grandis, the relative length of the caudal region is the same as in Pterothrissus, 
a discrepancy due to the majority of the caudal centra in /. grandis being 
shorter than the abdominal centra. /. macrocephalus, although having relatively 
fewer caudal vertebrae than I. grandis, shows a relatively longer caudal region than 
either I. grandis or Pterothrissus. The abdominal centra of /. macrocephalus are 
considerably shorter than the caudal centra, the reverse of the situation in /. grandis. 

The morphology of the neural arches and spines, the haemal arches and spines, 
the parapophyses, pleural ribs, intermusculars and supraneurals is similar to that 
described for Pterothrissus. In some specimens of Istieus irregularly shaped ridges 
are seen passing upwards and backwards from the bases of the neural arches. These 
ridges are interpreted as the impressions left by tough myocommata, a feature noted 
in Pterothrissus. 

Both the dorsal and anal fins are similar in shape, relative size and anatomy to 
those of Pterothrissus. The differences are merely those of meristic counts. 

The caudal skeleton of I. grandis is similar to Pterothrissus as may be seen by 
comparing Text-figs. 72 and 68. The important points of similarity are : the orna- 
mentation of the centra, the narrow neural and haemal spines, the two narrow uro- 
neurals, the narrow hypurals with a large gap between the second and third, the 



FOSSIL AND RECENT 

ep1-2 



147 



un2 




h3-6 



u2 



pu 5 



Fig. 72. Istieus grandis Agassiz. Caudal skeleton in left lateral view. 
Camera lucida drawing of B.M.N.H. 20586, fin-rays omitted. 



unexpanded bases of the inner fin-rays and the unsegmented condition of nearly all 
the basal fulcra. Dissimilarities are also evident : /. grandis shows only two epurals 
(the caudal skeleton of I. macrocephalus is not known in this respect), the second 
uroneural is relatively smaller, the first uroneural extends forwards to reach the 
second preural centrum and there are more basal fulcra. In the specimen figured 
(Text-fig. 72) hypural four is partially fused to the fifth hypural. 



Istieus macrocephalus Agassiz 
(Text-fig. 74) 

1833-44 Istieus macrocephalus Agassiz : 5, pt 1, p. 13, pt 2, p. 93, pi. 16. 
1863 ? Istieus macrocephalus Agassiz ; Marck : 39, pi. 4, fig. 6, pi. 5, fig. 3. 
1 90 1 Istieus macrocephalus Agassiz ; Woodward : 70. 

1954 1 stieus grandis Agassiz ; Siegfried : n. 

Diagnosis (emended). Istieus reaching 250 mm S.L. Operculum twice as deep 
as wide. Six branchiostegal rays. Vertebral column with not more than 75 
vertebrae. Dorsal fin with 41-44 rays. Body slender, with a narrow caudal 
peduncle. 



148 



ELOPIFORM FISHES 




T3 



a 
o 

ID 

A! 



a 

w 



be 
< 



s 

SS 

to 

s 







FOSSIL AND RECENT 149 

Holotype. B.M.N.H. P.3892, a nearly complete fish exhibiting a dorso-ventrally 
crushed cranium and lacking part of the tail. From the Campanian of Baumberg, 
Westphalia, Germany. 

Material. The holotype and B.M.N.H. 1275, a complete fish from the same 
locality as the holotype. The specimens are preserved in a buff-coloured limestone. 

Remarks. Agassiz (1835) separated /. macrocephalns from other species of the 
genus on the basis of its larger head. This difference is more apparent than real. 
The head length in the type-species and /. macrocephalus is similar, 30-34 per cent 
of the standard length. The maximum depth of the head is also comparable. Marck 
(1858, 1863) also recognized a species I. macrocephalus and followed Agassiz in quoting 
the size of the head as the diagnostic character, but parenthetically it should be added 
that the figure of I. macrocephalus given by Marck (1863 : pi. 4, fig. 6) shows a dorsal 
fin-ray count unlike that of Agassiz's type specimen. Thus, /. macrocephalus 
described by Marck may not be that of Agassiz (1835). 

Woodward (1901 : 71) recognized /. macrocephalus as a good species, but made no 
mention of head length. Instead he separated I. macrocephalus from the type-species 
on the basis that it had fewer dorsal and anal fin-rays, that the anal fin arises behind 
the dorsal and that the whole fish is of a more slender form. 

The dorsal fin-ray count is decidedly lower (41-44 in /. macrocephalus, 53-57 in 
/. grandis) and the fish is more slender, a feature probably giving the mistaken 
impression of a larger head. In /. macrocephalus the body tapers more markedly 
to a narrower caudal peduncle than in I. grandis. Contrary to Woodward's opinion 
(1901 : 71) the anal fin actually arises in the same position relative to the dorsal fin 
and vertebrae in both species, as may be seen by a comparison of Text-figs. 73 and 
74. The apparent posterior insertion of the anal fin in the type specimen of /. 
macrocephalus is due to the absence of the posterior dorsal fin-rays in the fossil. 

Other differences include : the vertebral count, 87-92 in /. grandis, not more 
than 75 in I. macrocephalus ; the operculum of /. macrocephalus is narrower (depth : 
width ratio of this bone is 2 : 1 against 1-4:1 in the type-species) and has a more 
oblique ventral margin ; there are only 6 branchiostegals in /. macrocephalus against 
8-10 in I. grandis. Finally, the snout appears more rounded in /. macrocephalus, 
but this may be due to distortion during preservation. 

Thus, /. macrocephalus may be separated from the type-species on the shape of the 
operculum, the number of branchiostegal rays, dorsal fin-ray and vertebral counts 
and the caudal peduncle depth. The general trend in Pterothrissidae appears to 
be towards elongation of both the vertebral column and dorsal fin. In this respect 
/. macrocephalus is more primitive than /. grandis ; in contrast, the lower number of 
branchiostegal rays would suggest that /. macrocephalus is more specialized than I. 
grandis. 

Remarks on other species 

Two other species of Istieus have been described, I. gracilis Agassiz from the 
Campanian of Sendenhorst and I. lebanonensis Davis from the Upper Santonian of 
Sahel Alma in the Lebanon. 



150 



ELOPIFORM FISHES 










c 
W 



< 












FOSSIL AND RECENT 151 

Istieus gracilis, so named because of the apparent slender form of this species, was 
considered by Woodward (1901) to represent specimens of I. grandis which had 
become distorted during preservation. I have examined neither the holotype nor 
many specimens which have been referred to this species. Two examples preserved 
in a single slab (B.M.N.H. P.3889), which were compared by Woodward (1901) to 
forms described as /. gracilis, show meristic counts similar to those of /. grandis. 
The evidence available to the author is not sufficient either to include the ' /. gracilis ' 
form with the type-species or to recognize a distinct species. Siegfried (1954 : 11) 
recognizes /. gracilis on the basis of the slender body form. 

Istieus lebanonensis is known only from the holotype, which is poorly preserved. 
Little may usefully be added to the descriptions given by Davis (1887) and Woodward 
(1901). The size and dorsal fin-ray count are closer to those of /. macrocephalus 
than /. grandis. 

Affinities of the genus Istieus 

Istieus was placed by Agassiz (1833-44) m the family Esocoides as the only 
marine genus in an otherwise freshwater group of fishes. In establishing the genus 
Agassiz (op. cit.) recognized four species, /. grandis, I. macrocephalus, I. gracilis and 
I. microcephalus. With the exception of /. macrocephalus the descriptions and figures 
given by Agassiz do not justify separation at the species level. 

Marck (1863) followed Agassiz, both with regard to familial placement and the 
recognition of several species, but he later (1873) combined /. grandis and /. micro- 
cephalus in a single species to which he gave the name /. macrocoelius. Such an 
action is contrary to the rules of nomenclature. Marck (1863) also established a new 
species, /. mesospondylus , which is here considered a synonym of the type-species. 

Woodward (1901) recognized two species from Westphalia (/. grandis and I. 
macrocephalus) and another from Sahel Alma (/. lebanonensis). These are the only 
species recognized here. Woodward (1901) placed the genus in the Albulidae to 
which it is clearly related. 

Istieus is very similar to the extant Pterothrissus, a view expressed by Woodward 
(1901) who stated that the two genera are not easily distinguished from one another. 
This similarity is endorsed by Greenwood et al. (1966) and Goody (1969b). Differ- 
ences that do exist between the genera are minor. Many have been mentioned 
above in the description of Istieus grandis. The more important differences are 
mainly in degree, not absolute : thus Istieus shows a larger head ; a wider skull 
roof ; a more terminal mouth ; a smaller first infraorbital which reaches well for- 
ward ; more extensive parasphenoid teeth ; more teeth on the endopterygoid and 
(particularly) the ectopterygoid ; and finally more teeth on the maxilla. In all 
these respects Istieus may be considered more primitive than Pterothrissus. Absolute 
differences between the genera are few. Istieus shows a difference in tooth size 
between those teeth borne by the upper and lower jaws, a well-developed ridge 
upon the metapterygoid, two epurals and a posteriorly situated coronoid process. 
In contrast Pterothrissus has premaxillary and dentary teeth of equal size, no well- 
defined ridge on the metapterygoid, three epurals and an anteriorly situated coronoid 
process. 



i 5 2 ELOPIFORM FISHES 

Except for the lower number of epurals and possibly the differentiation in tooth 
size, Istieus is the more primitive of the two genera. Both these features may not 
necessarily be of generic significance because the lower jaw dentition and the caudal 
skeleton of I. macrocephalus and I. lebanonensis are poorly known. 

Apart from the differences mentioned above there are morphometric and meristic 
dissimilarities, but there appears to be as much intrageneric variation within 
Istieus as there are intergeneric differences between it and Pterothrissus. 

The minor differences between Istieus and Pterothrissus are far outweighed by 
the many important features of similarity in neurocranial anatomy, the hyopalatine 
series, the infraorbital series, the dermal jaws, the form of the teeth, the cephalic 
sensory canal system and the fins (particularly details of the caudal skeleton). 

The great similarity in form leads one to ask whether it would be advisable to 
regard Istieus and Pterothrissus as congeneric. Such a question can only be answered 
subjectively. Considered alone, none of the differences mentioned above would 
separate these forms at the generic level, but the nature and sum of the differences 
lead me to suppose that Istieus is markedly more primitive, to a degree that can only 
be appreciated by conferring generic status on the Cretaceous forms. 

Genus HAJULIA Woodward, 1942 
1942 Hajulia Woodward : 557. 
Diagnosis. See Woodward (op. cit.). 
Type and only species. Hajulia multidens Woodward. 

Hajulia multidens Woodward 

1942 Hajulia multidens Woodward : 557, pi. 6, fig. 3. 
Diagnosis. See Woodward (op. cit.). 

Holotype. Nearly complete fish, no. 100509 in the American University, 
Beirut, from the Middle Cenomanian of Hajula, Lebanon. 

Remarks. Hajulia, from the Middle Cenomanian of Hajula, was considered an 
albuloid by Woodward (1942) who suggested a relationship with Istieus. 

The pterothrissid nature of Hajulia is indicated by the following features : the 
rounded trunk as evidenced by the presence of short, curved pleural ribs ; the shallow 
neurocranium which exhibits a slightly elongated snout ; the mouth which is slightly 
inferior, and the toothed maxilla which bears a single supramaxilla ; the palatal 
dentition in the form of large conical teeth on the parasphenoid, endopterygoid and 
possibly the basibranchium. The coronoid process is placed posteriorly, as in 
Istieus, and the quadrate/mandibular articulation lies beneath the posterior half of 
the orbit. There are less than 10 branchiostegals, although the exact number is 
not known. 

Between the occiput and the dorsal fin there is a complete series of slender supra- 
neurals. The dorsal fin itself is slightly elongated, situated in the centre of the back, 
and in profile resembles that of Istieus and Pterothrissus. 



FOSSIL AND RECENT 153 

The caudal skeleton has six slender hypurals and a half neural spine on the second 
preural centrum. The neural and haemal spines are not expanded, in contrast to 
those in albulids. Preceding the upper and lower principal rays there is, as in Istieus 
and Pterothrissus, a series of very slender basal fulcra. Two B.M.N.H. specimens 
(P. 13905 and P. 13906), although reasonably well preserved in the caudal region, 
show no caudal scutes as recorded by Woodward (1942). In contrast to other 
pterothrissids, there are three uroneurals. 

The vertebral column, with 50-55 vertebrae, is shorter than in other pterothrissids, 
but other features, although of a primitive nature, agree with this group. The 
anal fin is very short, consisting of only seven rays, unlike the longer anal fin of 
Istieus and Pterothrissus. 

The pectoral girdle of Hajulia, with the fin inserted just beneath the level of the 
vertebral column, is unlike that of the pterothrissids. In this, and the shape of 
the ventral part of the cleithrum, there is a resemblance to the tselfatioids (sensu 
Patterson 1967c), but there is nothing else about Hajulia to suggest such an affinity. 

The sum of these briefly noted morphological characters indicates that Woodward's 
(1942) suggested affinity of Hajulia was correct. However, incomplete knowledge 
of Hajulia precludes anything but a tentative assignment to the Pterothrissidae, 
and in consequence a fuller discussion of this genus is omitted. 

Family ALBULIDAE Bleeker, 1859 

Diagnosis. Albuloid fishes in which the snout is elongated, mouth inferior. 
Ethmoid commissure incomplete, running through premaxillae. Supraorbital 
sensory canal not extending onto parietal. Occipital condyle formed by basioccipital. 
Interorbital septum ossified. Lateral ethmoid sutured with broad parasphenoid. 
Parasphenoid and endopterygoid with a prominent crushing dentition opposed by 
similar dentition upon the basibranchial tooth plate. Basihyal tooth plate eden- 
tulous or absent. Vomer, dermopalatine, premaxilla and dentary with small needle- 
like teeth. Posterior infraorbitals broad, covering cheek. Supraorbital small. 
Hyomandibular-metapterygoid foramen present. Premaxilla tightly bound to 
mesethmoid ; maxilla moving independently, a single supramaxilla. Mandible 
with low coronoid process situated posteriorly ; articular and endosteal articular 
ossified separately. Branchiostegals more than 12 in number. Gular plate present. 
Dorsal and anal fins short based. Caudal skeleton with a full-length neural spine 
associated with the second preural centrum, two uroneurals. Inner caudal rays 
of each lobe without expanded bases. Caudal scute present above and below 
peduncle. 

Genus ALBULA Scopoli, 1777 

For synonymy see Hildebrand (1963). 

Diagnosis (emended). Albulid fish in which the mouth is decidedly inferior, 
the maxilla is edentulous and extends to the anterior margin of the eye. Quadrate/ 
mandibular articulation beneath anterior half of the orbit. Gular plate very small 



154 ELOPIFORM FISHES 

and disposed vertically. Vertebral column of 72-76 vertebrae. Caudal centra as 
long as deep and marked laterally by two grooves. Dorsal and anal fins without 
elongated terminal fin-ray. Caudal fin without fringing fulcra, preural neural 
spines expanded antero-posteriorly. Dorsal fin situated in the centre of the back. 
Anal fin nearer to the caudal peduncle than to the pelvic fin. 

Type-species. Esox vulpes Linnaeus. 



Albula vulpes (Linnaeus), 1758 
(Text-figs. 75-84) 

For synonymy see Hildebrand (1963). 

Diagnosis (emended). Albula in which the preorbital distance is considerably 
less than half the total neurocranial length. Quadrate/mandibular articulation 
beneath the lateral ethmoid. Parasphenoid, endopterygoid and basibranchial 
teeth of variable size, there being a decrease in average tooth size towards the edges 
of all tooth plates. Depth of neurocranium at the autosphenotic level equalling 
26-28 per cent of the neurocranial length. Sub-epiotic fossae without ridges along 
the medial borders. Operculum slightly deeper than wide. Cranial roofing bones 
smooth. Dorsal fin with 17-19 rays, anal with 8-9. 

Habitat. All tropical and subtropical seas ; in shallow water. Hildebrand 
(1963) gives a general account of Albula vulpes and refers to literature on the biology 
and ecology of the species. 

Remarks. Albula vulpes is often cited in literature. The descriptions of the 
cranium given by Ride wood (1904) and of the tail given by Hollister (1936) are 
particularly good. The comparative description given below is intended to supple- 
ment those works. 

Description. Neurocranium. The relative dimensions of the neurocranium 
are similar to those of Pterothrissus . Thus it is long and shallow, the snout is narrow, 
the otic bullae are large, the subtemporal and sub-epiotic fossae are deep and the 
small post-temporal fossae are inclined antero-medially. As in all albulids the 
neurocranium is more heavily ossified than in Pterothrissus. 

The skull roof (Text-fig. 75) shows several features commonly met with among 
albulids. The frontal is large with a deep trough anteriorly. Posteriorly the frontal 
reaches well back resulting in a small pterotic and parietal. The parietal is broader 
than long and of irregular shape. The nasal is large and in some specimens is 
ossified in two units, an anterior and posterior nasal ossicle. 

The dermethmoid is of a more complicated shape than it is in pterothrissids. 
The dermethmoid of A. vulpes is raised to form a median ridge which, as in ptero- 
thrissids, is braced anteriorly by inclined struts. Anteriorly the dermethmoid 
ridge is widest dorsally where it shows a well-marked central groove that is covered 
in life by taut skin. Posteriorly the walls of this groove form ridges which pass 
postero-laterally and ' fade out ' on the frontals. At the point where the ridges 



FOSSIL AND RECENT 



155 



SOSC 



. 10mm 



ot.s.c 




asp 



pto 



soc 



epo 



Fig. 75. Albula vulpes (Linnaeus). Neurocranium in dorsal view. Course of sensory 
canals shown by dotted line on left side. Composite of several B.M.N.H. specimens. 

converge there is a deep, anteriorly directed pit which ends blindly within the body 
of the dermethmoid ridge. The ethmoid cartilage is ossified at its extreme anterior 
end in older individuals where perichondral bone surrounds spongy endochondral bone. 
Anteriorly the mesethmoid bears two articulatory facets set one above the other. 
Both facets are for articulation with the premaxilla. Pterothrissus has a single 
premaxillary facet. 



156 



ELOPIFORM FISHES 

50C 



Ot.S.C 



Ptf 




sef 



boc 



5 mm 



Fig. 76. Albula vulpes (Linnaeus). Neurocranium in posterior view. 
Composite of several B.M.N.H. specimens. 



In posterior view (Text-fig. 76) the neurocranium shows large sub-epiotic fossae, 
small post-temporal fossae and large otic bullae. The supraoccipital is relatively 
larger than in Pterothrissus, due to its more extensive ossification. Two further 
features of interest in the posterior neurocranial wall of A. vulpes are the small 
' isolated ' portion of the pterotic and the foramen between the epiotic and supra- 
occipital. The ' isolated ' portion of the pterotic represents the wall of the caecal 
end of the subtemporal fossa. Such a condition of the pterotic appears to be unique 
among lower teleosts, but in these the subtemporal fossa is rarely as deep, and even 
more rarely is it associated with a deep sub-epiotic fossa. The foramen between the 
supraoccipital and epiotic is occupied by epaxial musculature passing through to the 
post-temporal fossa. 

The intercalar is small, as usual in albuloids, with little extension on the lateral 
face of the neurocranium. Posteriorly the surface of the intercalar bears ridges 
for the insertion of the tendon from the post-temporal. 

The lateral neurocranial wall has in part been described by Ridewood (1904). 
Significant points are the depth of the roofed dilatator fossa ; the posteriorly placed 
opening of the jugular canal, close to the suture between the prootic and exoccipital ; 



FOSSIL AND RECENT 



157 




T) 




<D 




A 






10 


•d 







H 


* - 


ro 


cu 




6£ 


tn 


cd 







c^ 


'-P 


H 


Ih 




O 


X 


en 


zi 


c 

HI 


£ 


co 

O 


pq 


P. 
CD 


E 


h 







M 


CD 


fe 


Ih 




rt 






CO 


T3 


C 





CD 


O 


| 


+j 


O 


cS 


O 


M 


Dh 




CO 




CD 


i 


£ 


V 







CO 



•S £ 





S « 




3 s-. 




•a ° 




2 § 

.2 




+-> 




V-< "S 




3 CO 




£ ft 




CD 




^ 




• +■> 








2 CO 

3 ■+■> 




« C 




OS CD 




C <o 




■ 3 *-> 




J fr 




d- CD 




Ih 






E 
E ■ 


^-1 







* 


-§ a 




-» c 




^C <d 




. CD 


J 


o-o 




*^.a 








£ '3 









HH 




to 



■58 



ELOPIFORM FISHES 



fapmx 



mm 




foa 



Fig. 78. Albula vulpes (Linnaeus). Neurocranium in ventral view. 
Composite of several B.M.N.H. specimens. 



FOSSIL AND RECENT 159 

hm 



pr m 




Fig. 79. Albula vulpes (Linnaeus). Hyopalatine series of left side in lateral view. 
Composite of several B.M.N.H. specimens. 

the development of a ridge running antero-ventrally across the otic bulla from beneath 
the hyomandibular foramen, providing an anchorage point for branchial muscula- 
ture ; and the deep excavation at the base of the ascending process of the para- 
sphenoid. This excavation is seen in other albuloid fishes although its significance 
is by no means clear. Postero-dorsally to the opening for the hyomandibular trunk 
there is often seen in large specimens (e.g. B.M.N.H. 83. 12. 15. 106) a foramen which 
leads directly to the cranial cavity. Since this foramen is not a constant feature 
it cannot be of any great significance. 

Within the orbit Albula shows a completely ossified interorbital septum which is 
sutured with a greatly expanded basisphenoid stem. Anteriorly the lateral eth- 
moid is large and stout, totally unlike that element in pterothrissids. Ventrally 
the lateral ethmoid is sutured with a thin lateral wing of the parasphenoid. The 
lateral ethmoid also bears a rounded articulatory area which articulates with the 
posterior process of the palatine. 

The parasphenoid is wide throughout much of its length and bears many hemi- 
spherical teeth. Each parasphenoid tooth is smooth. The tooth size varies con- 
siderably throughout the toothed area but generally those teeth towards the rear 
of the patch are substantially smaller than those anteriorly. The vomer bears 
three or four rows of needle-like teeth which are inserted on a short pedicel. 

Hyopalatine bones. The hyopalatine series is very much like that of Pterothrissus. 
The series is long and shallow ; the hyomandibular slopes forward and between it 
and the metapterygoid there is a large foramen ; the quadrate shows a posterior 
spinous process and the ventral margin lies almost horizontally ; the symplectic is 
flattened and lies against both the metapterygoid and quadrate ; the autopalatine 
is long and overlies much of the ectopterygoid. 

The palatine remains largely cartilaginous, except in larger individuals. Two 
processes of the palatine begin to ossify early ; anteriorly there is a rounded ethmoid 



i6o 



ELOPIFORM FISHES 

dsp 




sop 



10 mm 



Fig. 80. 



Albula vulpes (Linnaeus). Cranium in left lateral view. 
Composite of several B.M.N.H. specimens. 



process articulating with the mesethmoid ; posteriorly there is a large posterior 
(lateral ethmoid) process. The precocious ossification of these processes should 
not be interpreted as two embryonic divisions as Ridewood (1904 : 51) implied. 
Rather, these processes ossify early in response to a functional need. The dermo- 
palatine is larger than in pterothrissids and carries many more teeth. The ectoptery- 
goid is relatively large (cf . Pterothrissus) and bears a well-developed process upon its 
lateral surface. Ridewood (1904 : 51) reports the presence of a few ectopterygoid 
teeth but I have seen no example with teeth. The dentition of the endopterygoid 
is similar to that on the parasphenoid. 

Dermal upper jaw. The upper jaw is formed by the premaxilla and the maxilla, 
the latter taking little part in the functioning of the jaw. The premaxilla bears on 
its outer surface a series of struts which form an incomplete outer wall to the pre- 
maxillary sensory canal. Anteriorly the premaxilla articulates with the ethmoid 
region by two medially directed heads, one above the other. Such a double articula- 
tion is not found in other albuloids and probably arose by a subdivision of a single 
elongate facet like that in Pterothrissus. The double articulation makes union 
with the ethmoid very strong and virtually immovable. The oral border of the 
premaxilla bears a band of fine, pointed teeth. 

The edentulous maxilla is narrow anteriorly, but deepens posteriorly where it 
is partially overlapped by a single supramaxilla. The anterior end of the maxilla 
is turned inwards and is slightly swollen where it forms two rounded prominences. 



FOSSIL AND RECENT 161 

The dorsal prominence probably corresponds to the palatine projection seen in 
elopoids and osmeroidids ; the ventral projection corresponds to the maxillary head. 
The maxilla is capable of moving independently of the premaxilla. 

Mandible. The mandible is short, with a deep coronoid process situated pos- 
teriorly and a narrow symphysis. Both the endosteal articular and the sesamoid 
articular are large. The mandibular sensory canal is contained within a deep trough 
on the dentary and dermarticular. The dentition consists of a band of fine needle- 
like teeth on the dentary. 

Circumorbital series. The supraorbital and antorbital are slender and form a thin 
strut, as in Pterothrissus. Two rostral ossicles lie anterior to the first infraorbital 
and as in other elopiforms are barely larger than the contained sensory canal. 

The infraorbital series of Albula is typical of that of other albulids but differs 
somewhat from that series in the Pterothrissidae. The posterior infraorbitals are 
wide and completely cover the cheek region. The dermosphenotic is large, its 
posterior margin continuing that of the fifth infraorbital. As in Pterothrissus the 
infraorbital margin is rolled, but in A. vulpes the sensory canal is better protected 
by bone (cf . Text-figs. 80 and 66) . 

Cephalic sensory canal system. The system of canals upon the head is mainly 
contained in large troughs. The supraorbital canal runs within the frontal from the 
level of the autosphenotic to the level of the lateral ethmoid. Anterior to this the 
canal continues in the frontal trough and the nasal ossicle(s) to end blindly anteriorly. 
Gosline (1961) records a single median cavity separated from the supraorbital canal 
of either side by membrane. The epithelium lining this cavity appears to be supplied 
by a fine branch of the facial nerve and could therefore be related to the supraorbital 
canal. 

The otic division of the cephalic sensory canal runs within the pterotic parallel 
to its dorso-lateral edge. The canal opens posteriorly by two pores, one directed 
laterally and joining with the preopercular canal, the other directed posteriorly and 
receiving the supratemporal commissure and the lateral line. Innervation of the 
otic canal is from two sources : anteriorly the otic branch of the facial pierces 
the autosphenotic, and posteriorly a supratemporal branch of the vagus pierces the 
lateral wall of the dilatator fossa. 

The lateral portion of the supratemporal commissure is contained within the supra- 
temporal while the middle section lies in the skin. Neither the pterotic nor the 
parietal is grooved, as they are in Pterothrissus. 

Hyoid arch, gill arches and gular plate. The ceratohyal consists of a stout posterior 
element and a longer but narrower anterior element. The latter bears a deep groove 
on the lateral surface for the reception of branchiostegal rays. The afferent hyoidean 
artery does not pierce either hypohyal ; instead it passes upwards within the cerato- 
hyal and runs back along a groove on the dorsal edge of the anterior ceratohyal. Of 
the 15 branchiostegal rays, the anterior three lie free in the branchiostegal membrane. 

The gill arches are very similar to those of Pterothrissus. Minor differences that 
do exist concern the teeth borne by the basibranchial tooth plate. In Albula vulpes 
these teeth are hemispherical, irregular in size and show no pattern of size distribu- 
tion. A basihyal tooth plate is absent. 



1 62 



ELOPIFORM FISHES 




^r 



1) (D « 



3 b 






2 O 

G S d 

u) +J O 

■3 «" "■fi 

ro o ri 

.15 „ d 



U r* J-* 

.S ^3 +j 

H (11 in 



MH > 



T3 

d 

0) 



x> d 
%>2 



s d 



O io 
" O 

On 



az 



M s 

■ 5 C •"" 

o .* S ft 



FOSSIL AND RECENT 



163 



stt. com 



stt 




pel 1-3 



cor 

Fig. 82. Albula vulpes (Linnaeus). Pectoral girdle of right side in medial view. 
Composite of several B.M.N. H. specimens. 



A small gular plate is seen in some specimens (Nybelin 1960), notably the larger 
individuals. Unlike the gular plate of other albuloids that of A. vulpes is vertical. 

Paired fins. All bones composing the pectoral girdle (Text-fig. 82) are stout. The 
main body of the post-temporal is longer than wide and is produced into three pro- 
cesses, an epiotic (dorsal) limb, an intercalar (ventral) limb and a third limb which 
projects into the lumen of the post-temporal fossa and is embedded within epaxial 
musculature. The supracleithrum is the usual elongate element. 

Many features of the ventral parts of the pectoral girdle are similar to those of 
Pterothrissus. The cleithrum is broad above the fin insertion but becomes narrow 
antero-ventrally ; the coracoid does not extend to the tip of the cleithrum ; an 
interosseous foramen is absent (in Pterothrissus it is very small) ; the scapula fails 



i6 4 ELOPIFORM FISHES 

completely to enclose the scapular foramen ; and, finally, the anterior edge of the 
mesocoracoid fits tightly against the inflected margin of the cleithrum. The ventral 
end of the mesocoracoid is bifurcated, producing a foramen between the mesocoracoid 
and the coracoid. As in other parts of the skeleton the pectoral girdle of A . vulfies 
is more extensively ossified than that of Pterothrissus. 

There are three postcleithra, the lowermost having a characteristic shape (Text- 
fig. 82). 

The disposition, size and shape of the pectoral radials is similar to that described 
for Pterothrissus. There are 16-17 pectoral fin-rays of which all except the outermost 
are branched. A triangular pectoral splint is associated with the outermost ray. 

The pelvic fin arises beneath the posterior half of the dorsal fin and is supported 
by a triangular pelvic bone. Anteriorly the pelvic bone is thin, but posteriorly it 
becomes considerably thicker and the hind margin is capped by cartilage with which 
the pelvic rays articulate. Medially, the base of the pelvic bone is produced as an 
' ischial process '. There are 10 pelvic fin-rays, the outermost unbranched. As- 
sociated with the upper half of the first ray is a long splint bone, the proximal end 
of which is curved sharply upwards. Whitehead (1963 : 744, fig. 3a) shows the pelvic 
splint to be a double structure composed of an upper and lower portion. I could 
not find any convincing example of such a double structure in any B.M.N.H. 
specimen. 

Median fins. The dorsal is situated nearer to the occiput than to the caudal 
peduncle and is composed of 17-18 rays, the anterior five of which are unbranched. 
The fin is supported by 15-16 pterygiophores. In older specimens successive ptery- 
giophores appear to be fused distally. 

The small anal fin is remote and composed of nine rays (occasionally eight) 
supported upon seven pterygiophores. Like the pterygiophores of the dorsal fin, 
the distal extremities may become fused in older specimens. 

Vertebral column. The only complete vertebral count available to me (B.M.N.H. 
83. 12. 15. 106) showed 67 vertebrae of which 23 are caudal. Hildebrand (1963 : 134) 
gives the range of vertebral counts for the western North Atlantic forms as 72-74. 
All centra except the first 17 are as deep as long, the exceptional centra being deeper 
than long. Each centrum is marked laterally by a stout horizontal bar flanked by 
deep grooves. The centra bear deep pits dorsally and ventrally for the reception of 
neural arches and parapophyses or haemal arches respectively. 

The bases of the neural arches are produced anteriorly and posteriorly forming 
small zygapophyses. The first 31 neural arches and spines remain distinct from their 
fellows of the opposite side and bear separate neural spines. Thus these vertebrae 
show a double neural spine. 

The parapophyses of the first 44 centra show an elongate area which is inclined 
antero-ventrally and serves, with a facet on the centrum, for the articulation of the 
pleural rib. The posterior 12 parapophyses gradually increase in length and support 
a series of progressively shorter pleural ribs. 

The neural and haemal spines of the caudal region are short. Those associated 
with the posterior centra are expanded in the sagittal plane and aid in supporting 
the caudal fin-rays. Preceding the first dorsal pterygiophore there is a series of 



FOSSIL AND RECENT 



165 



ep1-2 un2 




h1-2 



Fig. 83. 



Albula vulpes (Linnaeus). Caudal skeleton in left lateral view. Arrows indicate 
upper and lower principal fin-rays. Composite of several specimens. 



supraneurals. Each is sigmoid in shape, with the anterior ones considerably stouter 
than the posterior members of the series. 

Epineural intermuscular bones are borne by the neural arches throughout most of 
the abdominal and caudal regions. Epipleurals are found in the posterior abdominal 
region and grade posteriorly into a ventral series found throughout the caudal region. 

Squamation. Lateral line scale counts vary considerably. Specimens from both 
coasts of Central America give counts ranging from 71-77. Hildebrand (1963 : 134), 
who based his counts on specimens from the western North Atlantic, records a range 
of 65-67. It may be mentioned here that an Aden specimen (B.M.N.H. 1962.3.26.9) 
also differs from the western Atlantic and eastern Pacific specimens in having 19 
dorsal fin-rays. Hildebrand (1963) observed that only occasionally do Atlantic 
specimens show such a high count. Regrettably the dorsal fin of a Seychelles speci- 
men was broken and accurate counts could not be made. 

The branchiostegal ray count also varies with respect to geographical distribution 
(McAllister 1968 : 37) being higher in fishes from India and the Indo-Australian 
Archipelago. These biometrical differences may justify the recognition of a second 
Recent species. 

Returning to squamation, the transverse count immediately anterior to the dorsal 
fin is eight above and nine below the lateral line row, excluding a median row of small 



1 66 



ELOPIFORM FISHES 




30 mm 



Fig. 84. Albula vulpes (Linnaeus). Entire skeleton, scales omitted. 



scales which extends from the occiput to just behind the dorsal fin. The general 
body squamation consists of thin cycloid scales, more or less square in shape, with 
three to four basal radii and four to five basal lobes. The posterior edge of the 
exposed field is often broken due to the thinness of the scale. The surface of the 
scale is marked only by circuli. 

The modified scale row in the mid-dorsal line consists of oval scales, whose length 
considerably exceeds their width. The bases of all fins are covered by scales. 
Axillary scales are present at the bases of the pectoral and pelvic fins. Bone cells 
are present in the scales. 



Albula oweni (Owen) 

(Text-figs. 85-87) 

1840 Pisodus owenii Owen : 138, pi. 47, fig. 3. 

1844 Pisodus owenii Owen ; Agassiz : 2, pt 2, p. 247 (name only). 

1854 Pisodus owenii Owen ; Owen : 167. 

1891 Pisodus oweni Owen ; Woodward : 108, pi. 3, figs. 3-5. 

1893 Pisodus oweni Owen ; Woodward : 357, pi. 17. 

1901 Albula oweni (Owen) Woodward : 60, pi. 4, figs. 1, 2. 

1908 ? Albula oweni (Owen) ; Priem : 82, fig. 38, pi. 4, fig. 12. 

1946 Albula oweni (Owen) ; Casier : 122, pi. 3, fig. 12. 

1952 Albula oweni (Owen) ; Arambourg : 242, pi. 37, figs. 1-6. 

1966 Albula oweni (Owen) ; Casier : 133, pi. 13, fig. 4. 

Diagnosis (emended). Albula in which the preorbital part of the neurocranium 
is equal to half the total neurocranial length. Quadrate/mandibular articulation in 
advance of the level of the lateral ethmoid. Parasphenoid teeth of the same size 
throughout the toothed area. Depth of the neurocranium at the autosphenotic 
approximately 35 per cent of the neurocranial length. Opercular bone twice as 
deep as wide. Ornamentation in the form of coarse rugosities present on the lateral 
edges of the frontals above the orbit and on the quadrate. 



FOSSIL AND RECENT 



167 



SO. SC 



Ot.S.C 




asp 



soc 



Fig. 85. Albula oweni (Owen). Neurocranium in dorsal view. Course of sensory canals 
shown by dotted line on left side. Based on B.M.N.H. 39439. 



1 68 



ELOPIFORM FISHES 

soc 



ot s.c 




exo 



Fig. 86. Albula oweni (Owen). Neurocranium, posterior view of dorsal half. 

From B.M.N. Ff. 39439. 

Holotype. Parasphenoid ; in the collection of the Royal College of Surgeons, 
London. From the London Clay (Ypresian) of Sheppey, Kent, England. 

Material. B.M.N.H. 39439, P.9158, and various isolated teeth, teeth with sup- 
porting bones, and a decaying specimen of the viscerocranium, all from the London 
Clay (Ypresian) of Sheppey. Postcranial remains are unknown. 

Horizon and locality. Lower Eocene of S.E. England, France, Belgium and 
N. Africa, and Middle Eocene of Belgium. 

Descriptive remarks. Woodward (1901) distinguished A. oweni from A. vulpes 
by the relative lengths of snout to frontal notch in the two species. The frontal 
notch in Albula occurs at the hind margin of the lateral ethmoid and is thus compar- 
able with the preorbital distance, a feature used here. 

Several other neurocranial features distinguish A. oweni from the type-species. 

In dorsal view the skull roof of A. oweni shows a marked median frontal depression 
at the level of the autosphenotic spine. Such a depression is suggested in A. 
vulpes, but it never attains the same relative proportions. Perhaps its development 
is a consequence of absolute size, this being greater in A. oweni. Above the orbit, 
the frontals exhibit a rugose ornamentation not seen in the extant species. Such 
ornament is seen in the Osmeroididae where it is often (Osmeroides lewesiensis) more 
extensive. Apart from these minor differences the skull roof of A. oweni (Text-fig. 
85) is very similar to that of A. vulpes (Text-fig. 75). 

The posterior neurocranial wall shows well-developed sub-epiotic fossae (an 
albuloid feature) which are bordered medially by stout vertical ridges (a feature 
typical of the osmeroidids but not found in A. vulpes). The supraoccipital of A. 
oweni (Text-fig. 86) is of different proportions to that of A. vulpes and there is a small 
interspace between the supraoccipital and exoccipitals. Neither an ' isolated 



FOSSIL AND RECENT 



169 



portion ' of the pterotic (p. 156) nor a foramen between the supraoccipital and 
epiotic could be identified in A . oweni. 

The depth of the neurocranium at the autosphenotic spine is relatively greater in 
A . oweni than in A . vulpes. The lateral neurocranial wall of A . oweni is incompletely 
known from a single specimen, and comparison with the extant species is difficult 
as features of individual variation cannot be taken into account. However, a few 
remarks are necessary. 

The region which is occupied by the saccular swelling in A. vulpes is depressed 
and perhaps represents the point to which an arm of the swimbladder reached. In 
A. vulpes the swimbladder diverticulum, although of variable development (Green- 
wood 1970a), never reaches as far forwards as this point. 

The subtemporal fossa of A . oweni is considerably larger than in A . vulpes and a 
horizontal ridge forms its ventral border. In A. vulpes no such ridge exists. 

The posterior opening of the jugular canal (pars jugularis) lies, as usual, within 
the prootic bone, but in A. oweni the opening is situated far forwards, whereas in 
A. vulpes it is close to the posterior margin of the prootic. The specimen of A. oweni 
shows a lateral bridge of bone spanning the jugular groove behind the posterior 
opening of the pars jugularis. A greater anterior development of this lateral bridge 
would eventually meet the lateral commissure and produce the same spatial rela- 
tionship of the posterior opening of the jugular canal as in A. vulpes. This develop- 
ment of a bridge, which may otherwise be thought of as a fenestrated lateral 
commissure, is in all probability subject to individual variation. 

Lying antero- ventral to the posterior opening of the pars jugularis is a large 
postero-ventrally directed boss ; in the type-species there is only a stout ridge. The 



25mm 




foa 



par 



Fig. 87. Albula oweni (Owen). Neurocranium in right lateral view. 
Based on B.M.N. H. P. 9158 and 39439. 



170 ELOPIFORM FISHES 

tip of this boss in the prepared specimen (B.M.N.H. 39439) is spongy, suggesting that 
it was capped by cartilage. This boss probably received branchial musculature, 
as does the corresponding ridge in A. vulpes. 

The dentition on the parasphenoid and vomer is very similar in both species, yet 
there are small differences. The parasphenoid teeth of A. vulpes show a gradation 
in average size from large anteriorly to smaller posteriorly where they also become 
more conical in shape. Parasphenoid teeth in A. oweni are constant in shape and 
size throughout their extent. The vomer of both species bears a small, raised, 
transversely orientated tooth patch of small needle-like teeth. A specimen of A. 
oweni (B.M.N.H. 30528) shows this tooth patch to be partially subdivided into right 
and left halves (Woodward 1901 : pi. 4, fig. 2) suggesting the origin of the undivided 
condition in A . vulpes from paired elements. 

The quadrate/mandibular articulation of A. oweni lies relatively further forward 
than in the type-species. The lower jaw is shorter, a feature accentuated by the 
relatively longer snout. A consequence of the greater suspensorial angle (Gregory 
1933) is that the horizontal limb of the preoperculum is longer than the vertical limb. 
In A . vulpes the limbs are of equal length. 

The differences mentioned above are minor and serve only to distinguish A . oweni 
as a distinct, but closely related species. To determine which is the more primitive 
is not easy. There are several differences which appear to have little or no sig- 
nificance in this case, such as the median frontal depression, deeper neurocranium 
and shorter jugular canal in A. oweni. 

A. oweni appears more specialized than A. vulpes in having a relatively longer 
snout and a more forwardly positioned quadrate/mandibular articulation. In 
other features, however, there is a greater resemblance to the primitive Osmeroides, 
for example, the ornamentation, the ridge at the medial margin of the sub-epiotic 
fossa and the partially divided vomerine tooth patch. 

The majority of specimens of A. oweni are isolated teeth or fragments of tooth 
plates, making specific determination difficult, particularly since the teeth associated 
with the crania described above are poorly preserved. However, if B.M.N.H. 
P. 1697 (a group of teeth from the Lower Hamstead Beds of the Isle of Wight) is 
correctly identified, then this species must have extended to the Lower Oligocene 
(Lattorfian). 



Albula eppsi White & Frost 

1931 Albula eppsi White & Frost : 83-84, 105, figs. 137-14, pi. fig. 8. 
i960 Albula eppsi White & Frost ; White : pi. 30, fig. 4. 

Diagnosis. See White & Frost 1931 : 83. 

Syntypes. B.M.N.H. P.14628, P.i4628a-b, tooth plates from the Blackheath 
Beds (Sparnacian), Abbey Wood, Kent, and P.15287 from the same horizon and 
locality. 

Material. Twenty-three specimens in the B.M.N.H. consisting of fragmentary 
tooth plates and a small portion of an operculum. 



FOSSIL AND RECENT 171 

Horizon and locality. Sparnacian-Cuisian, S.E. England. 

Remarks. White & Frost (1931) stated that the tooth plates referred to A. eppsi 
differ from those of the Recent species in showing teeth of equal size, with no diminu- 
tion of tooth size towards the margin of the tooth plate. Such a regular tooth size 
is characteristic of the contemporaneous A . oweni and this fact probably motivated 
Casier (1966 : 134) to interpret A. eppsi as representing young individuals of A. 
oweni. However, the great difference in size between the two species and the total 
absence of intermediate forms suggests that this species is probably a valid one. 
A . eppsi is reminiscent of Osmeroides latifrons (Turonian) in being small and showing 
no diminution of tooth size towards the edge of the tooth plate. 

Frost (in White & Frost 1931 : 105) referred an otolith to A. eppsi, presumably 
because it was found together with tooth plates of this species. This otolith is held 
by Frost to resemble those of A. vulpes in the rounded anterior and posterior rims. 

Other albulid remains 

There is in the B.M.N.H. collection a large series of isolated teeth from Tertiary 
deposits of Europe which, although albulid in form, are not easily referred to any 
of the species recognized so far. Until more associated material of various growth 
stages become available the specific identity of these teeth must remain in limbo. 

Woodward (1901 : 73) referred to several specimens from the Gault Clay (Albian) 
which, he said, may be albulid. The specimens concerned (B.M.N.H. 47286, 
P. 9159) are phyllodont tooth plates and should be referred to Casierus Estes or a 
closely related form. 

In the B.M.N.H. collection there are two specimens, B.M.N.H. P. 1224 and 
P.1225, which are albulid in form and come from the Upper Cretaceous of Red Deer 
River, Alberta, Canada. The specimens consist of isolated teeth and vertebrae. 
The teeth are unlike those of any Albula species in being laterally flattened and in 
showing a narrower base than crown. The vertebrae, which in lateral view are 
deeper than wide, are marked by numerous fine anastomosing strengthening bars. 

There are records of other albulids from North America. Applegate (1970) has 
described a new species of Albula, A. dunklei, from the Moore ville Chalk (Santonian) 
of Alabama. Estes (1969c) erected a new monotypic albulid genus Coriops amnicolus 
for specimens from the Lance Formation of Wyoming and the Hell Creek Formation 
of Montana, the age of which, according to Cobban & Reeside (1962), is Maastrich- 
tian-Danian. The albulid specimens found in Alberta, Wyoming and Montana are 
from freshwater deposits. 

Other Upper Cretaceous albulids of N. America have been recognized on scale 
remains by Cockerell (1933), who described A. antiqua from Florida, and David 
(1940) who described a different genus, Kleinpellia. 

Genus LEBONICHTHYS nov. 

Diagnosis. Albulid fish in which the frontals and pterotics show ornamentation 
in the form of weak ridges. Mouth terminal or nearly so. Maxilla extending to 
beneath the posterior half of the orbit, with or without teeth. Single elongate 



172 ELOPIFORM FISHES 

supramaxilla. Quadrate/mandibular articulation beneath posterior half of the orbit. 
Suboperculum large, equal in area to half that of the operculum. Gular plate large. 
Vertebral column with 64-66 vertebrae of which 23-24 are caudal. Centra deeper 
than long, marked laterally by 5-6 grooves. Anal fin situated nearer to caudal 
peduncle than to the pelvic fin. 

Type species. Osmeroides gracilis Davis. 



Lebonichthys gracilis (Davis) 

(Text-figs. 88-91) 

1887 Osmeroides gracilis Davis : 559, pi. 31, figs. 2, 3. 
1887 Osmeroides latus Davis : 561, pi. 31, fig. 1. 
? 1887 Osmeroides brevis Davis : 560, pi. 32, fig. 1. 
1901 Osmeroides gracilis Davis ; Woodward : 16, pi. 2, figs. 1, 2. 

Diagnosis. Lebonichthys reaching 170 mm S.L. Head and opercular apparatus 
equal in length to 30 per cent of standard length. Parasphenoid with hemispherical 
teeth of irregular size. Ectopterygoid edentulous. Maxilla edentulous. Dorsal 
fin situated nearer to the occiput than to the caudal peduncle, with 24-25 rays. 
Caudal skeleton with three epurals, supporting neural spines unexpanded, fringing 
fulcra and caudal scutes present. Scales with anterior margin scalloped. 

Holotype. Incomplete fish, R.S.M. 1891.59.86, from the Upper Santonian of 
Sahel Alma, Lebanon. 

Material. The holotype and 16 specimens in B.M.N.H. All specimens are 
from the Upper Santonian of Sahel Alma, Lebanon. 

Description. The head occupies 30 per cent of the standard length and is twice 
as long as it is deep. The anterior end of the snout is slightly truncated. The 
quadrate/mandibular articulation lies beneath the centre of the orbit. With the 
exception of the frontals and pterotics, the cranial bones are smooth. 

Neurocranium. The neurocranium is shallow throughout. The roof is flat with 
the frontals forming the majority of its area. Each frontal meets its partner in a 
slightly wavy suture. Anteriorly the dermethmoid, which is smaller and simpler 
in shape than that of Albula, separates the frontals. The frontal tapers gradually 
toward the snout but there are two places where the frontal narrows sharply, at the 
level of the autosphenotic spine and again at the level of the lateral ethmoid. Pos- 
teriorly the frontal is indented where it meets the pterotic. The surface of the frontal 
is slightly raised at the centre of ossification. From this centre several ridges, 
with intervening grooves, curve anteriorly and laterally. The most medial of these 
ridges ends anteriorly in a foramen which marks the point of exit from the bone of 
the supraorbital sensory canal. Anterior to this level the canal continued in a 
shallow groove. The central portion of the skull roof exhibits a slightly depressed 
area. Faint ridges may also be seen passing posteriorly from the ossification centre 
of the frontal but these are weak compared to the anterior ridges. 



FOSSIL AND RECENT 



173 



SO.S.C 




epo 



soc 



Fig. 88. Lebonichthys gracilis (Davis). Neurocranium in dorsal view. 
Based on R.S.M. 1891.59.85 and 1881.5.36. 



The parietal is rhomboidal and meets its partner in a nearly straight median suture. 
Between them the parietals cover much of the supraoccipital and the epiotics, only 
the crest of the former and the processes of the latter being visible in dorsal view. 

The pterotic meets the frontal in a syndesmotic union, the path of the suture 
line being due to the posterior margin of the frontal rather than the anterior border 
of the pterotic which lies beneath the frontal. The posterior margin of the pterotic 
is oblique and meets the lateral margin in a small laterally directed process which 
marks the posterior end of the hyomandibular facet. The surface of the pterotic 
is marked by fine ridges extending forwards from the centre of ossification. The 
otic sensory canal ran in a deep groove incompletely roofed by bone. Posteriorly 
the canal opened by a single pore at the level of the lateral prominence on the margin 
of the pterotic. 

The anterior part of the neurocranium is defective in all specimens. The derm- 
ethmoid is incompletely known but a median vertical ridge may be seen in several 



174 ELOPIFORM FISHES 

examples. The nasal is a thin, gutter-shaped element partially overlapping the 
anterior end of the frontal. 

Of the lateral neurocranial wall little can be seen. Features which may be recog- 
nized are a deep subtemporal fossa beneath the hyomandibular facet, a large otic 
bulla which accommodated the sacculith, and a very small autosphenotic spine. 
Imperfect suture lines between the various neurocranial elements are seen in some 
specimens. The outline and contours of those which may be identified suggest that 
the neurocranium is very much like that of Osmeroides latifrons. Significantly, 
the intercalar is reduced to a small cap of bone postero-ventral to the subtemporal 
fossa. 

The parasphenoid is distinctive in bearing a dentition similar to that of small 
specimens of Albula vulpes. The bone is broad throughout its length. Posteriorly 
it reaches the occipital condyle, while anteriorly it overlies the vomer beneath the 
lateral ethmoid. At the level of the weakly defined ascending wings the para- 
sphenoid exhibits a shallow angle. Posterior to this level the parasphenoid extends 
back and terminates in two short prongs, between which there is a median ventral 
groove extending forwards to the parasphenoid flexure. Beneath the orbit and part 
of the lateral ethmoid the parasphenoid bears a tooth plate. Each of the many 
teeth borne by this plate is hemispherical and bears a distinct translucent apical 
cap of enamel. There is a large pulp cavity. The posterior teeth are the smallest, 
the largest being found in the anterior third of the plate. Throughout the toothed 
area very small teeth may be seen lying adjacent to large ones. Towards the lateral 
margin of the plate there is a predominance of smaller teeth. The base of the 
ascending wing of the parasphenoid is marked by a deep excavation at the bottom 
of which lies the foramen for the internal carotid artery. The efferent pseudo- 
branchial artery must have looped around the antero-dorsal margin of the para- 
sphenoid ascending wing. 

The vomer is as wide as the anterior end of the parasphenoid, with which it is in 
contact in an overlapping suture. No specimen shows definite evidence of vomerine 
teeth although several small needle-like teeth, which may probably be referred to the 
vomer, are often found in the snout region. 

Hyopalatine bones. The hyomandibular is stout and relatively short. Dorsally 
the bone is thick and forms a single broad articular head. The bone projects an- 
teriorly as a thin wing which meets the metapterygoid. Characteristically, there is 
little or no overlap between these elements. The ventral part of the hyomandibular 
slopes forwards as a stout shaft, the posterior margin of which is grooved for the 
passage of the hyomandibular trunk of VII and the efferent hyoidean artery. The 
opercular process is weakly defined. 

The metapterygoid is of irregular shape. Anteriorly this element overlaps the 
endopterygoid while ventrally it is separated from the quadrate by a narrow inter- 
space. The postero-ventral margin of the metapterygoid overlaps the symplectic, 
as in other albuloids. 

The symplectic is long in Lebonichthys and provides a firm linkage between the 
forwardly positioned quadrate and the ventral end of the hyomandibular. Pos- 
teriorly the symplectic is somewhat flattened. 



FOSSIL AND RECENT 



175 




E 
E 



CJ 




G 




O 




-O 




<D 




.c 








'-P 




cti 








13 




ft 







>> 


-a 






co 




C 


+-> 


0) 


£ 


.1 


b£ 







S 


CD 

ft 




CO 




'> 


£ 


"c3 


a 


l-C 

c 


pq 


CD 
> 


t-, 


_fl 


> 

CD 


T) 


CO 




C(H 


O 


O 


d 




<u 


CO 


A 


CD 
4J 


ft 

co 


'to 


ft 


ft 


5 








-P 




*4H 




CD 
h-1 


CD 








> 


"ST 


2 


a 

Q 


3 
S 


to 


g 


"*■» - 








"8 -S 

e 5 


60 


CO 


+-> 


00 •*-( 


£' 










00 







176 ELOPIFORM FISHES 

The quadrate is fan-shaped. The posterior or ventral margin lies almost hori- 
zontally, with the expanded condyle directed antero-ventrally. A groove on the 
medial surface of the quadrate received the lower end of the symplectic. 

The postero-ventrally inclined limb of the ectopterygoid fits tightly around the 
anterior border of the quadrate. The longer antero-dorsally inclined limb runs 
along the entire length of the endopterygoid. A shallow crest of bone arises from 
the dorsal margin of the ectopterygoid at the angle in the bone. This crest is similar 
to that developed in Osmeroides latifrons and homologous with the peculiarly shaped 
ectopterygoid process of Recent albuloids. The endopterygoid is oval, at least three 
times as long as wide. The convex buccal surface bears a large patch of teeth of the 
same form but of smaller average size than those borne by the parasphenoid. The 
anterior limit of both endopterygoid and ectopterygoid is difficult to decipher in 
any specimen as this region is covered by the crushed remains of a spongy ossification 
representing a weakly ossified autopalatine. The autopalatine of Recent albulids is 
only weakly ossified in large individuals. The anterior end of the autopalatine in 
Lebonichthys is represented by a rounded knob formed of more compact bone than 
that posteriorly. The palatine head articulates with the mesethmoid medially and 
the inner face of the maxillary head anteriorly. B.M.N.H. P. 4855 shows a small 
patch of palatine teeth which are villiform in shape, not needle-like as in Albula. 

Dermal upper jaw. Those specimens exhibiting a premaxilla indicate that there 
was a very close union between this element and the ethmoid. The premaxilla is 
broad anteriorly where it meets its partner in the mid-line. Posteriorly the alveolar 
portion is tapered. The lateral surface of the premaxilla is marked by two or three 
pores but interconnecting canals cannot be traced ; thus the presence of a premaxil- 
lary sensory canal is known by inference rather than from direct evidence. A narrow 
band of fine, needle-like premaxillary teeth is found along the slightly convex oral 
border. 

Unlike the premaxilla, the maxilla is often found dissociated from the neuro- 
cranium. Anteriorly the expanded head of the maxilla turns dorsally and medially. 
Posteriorly the maxilla expands into a shallow ovoid which is overlain postero- 
dorsally by a single elongate supramaxilla. The maxilla is grooved dorsally behind 
the maxillary head and ventrally beneath the head. The dorsal groove marks the 
position occupied by the overlying first infraorbital. The ventral groove contained 
the alveolar portion of the premaxilla. The maxilla is edentulous. 

Mandible. The mandible is short and bears a prominent coronoid process deve- 
loped towards the posterior end of the jaw. The dentary symphysis is turned 
slightly ventrally. Distinctive albuloid features of the mandible are the inflected 
ventral margin and the presence of a deep groove containing the mandibular sensory 
canal. The mandibular sensory canal was not covered by bone ventrally. The 
dentary bears a band of needle-like teeth along the oral margin. Posteriorly the 
articular forms the articulatory cup. The medial surface of the mandible bears a 
large endosteal articular and sesamoid bone. 

Hyoid arch, branchiostegal rays, gular plate and gill arches. The ceratohyal is, as 
usual in teleosts, ossified in two sections, a large anterior and small posterior element. 
The former is represented by a thick bone which is constricted in the anterior third. 



FOSSIL AND RECENT 177 

The anterior end is shallower than the posterior end. The posterior ceratohyal is 
thinner than the anterior element. The ventral margin of the posterior ceratohyal 
is curved while the dorsal margin is straight, except posteriorly where the border 
is notched for the reception of the interhyal. Both dorsal and ventral hypohyals 
may be identified but little of their shape may be seen. Each hypohyal consists 
of a core of spongy bone covered by a thin perichondral shell. The anterior tip of 
the hypohyals lies between and level with the lower jaw articulation when the 
mouth is closed. The dorsal edge of the anterior ceratohyal is grooved for the afferent 
hyoidean artery. The depth of this groove is comparable with that seen in Albula. 

There are 12-14 branchiostegal rays. The first (or most posteriorly situated) 
branchiostegal is broad and supported solely by the posterior ceratohyal. The 
second of the series articulates with the hyoid bar at the junction of the two cerato- 
hyals. The branchiostegals become progressively thinner at the anterior end of the 
series. Anteriorly, the last ray in the series may have lain free in the branchio- 
stegal membrane since it is often displaced and shows no development of an articu- 
latory head. 

A large gular plate is attached to the posterior edge of the dentary symphysis. 
The width of this plate, equivalent to over one-third of the mandibular length, 
suggests that the lower jaw rami were wide apart. It is significant that in many 
specimens the jaw rami are often seen splayed apart. The size of the gular plate 
would preclude any interpretation of the gular lying vertically as it does in the 
Recent Albula. 

The basibranchial and basihyal elements are poorly known. A large basibranchial 
tooth plate lies in a position (with respect to the hypobranchials) suggesting that it 
represents the fusion of the first three basibranchial plates. The dorsal surface of 
this plate bears many villiform teeth. The posterior end of a basihyal tooth plate 
is seen in B.M.N.H. P. 4764. The plate appears edentulous. 

The hypobranchials of the first three arches are developed in typical fashion. The 
first is the longest of the series and shows a slight curvature. The second, like the 
first, shows a distinct basal articulatory head proximally, but the short third hypo- 
branchial appears merely as an ' outpushing ' from the third basibranchial. The 
five ceratobranchials are typical in that the first is substantially longer and broader 
than the fifth. The latter elements share a small mutual interface. Small, irregular 
tooth plates, each with many needle-like teeth, are associated with the oral surfaces 
of the ceratobranchials. These tooth plates are relatively sparse : for instance 
six are detectable upon the fourth ceratobranchial of B.M.N.H. P. 4764. A single 
large tooth plate is associated with the medial margin of the fifth ceratobranchial. 
This tooth plate, the lower pharyngeal, has the shape of an isosceles triangle, twice 
as tall as wide with the apex directed anteriorly. 

Well-defined gill-rakers were not seen on any but the first ceratobranchial. The 
gill-rakers have the same form as those of Albula and Pterothrissus. Nothing of the 
dorsal gill arch elements was seen. 

Circumorbital bones. The supraorbital was not positively identified in any speci- 
men but several exhibit small splinters of bone indicating that its form was as in 
Recent albuloids. 



178 ELOPIFORM FISHES 

Similarly, the infraorbitals are incompletely known. The orbital margin of this 
series is gently rounded, as in Albula but in contrast to Pterothrissus. Anteriorly, 
the dorsal margin of the first infraorbital is straight and inclined towards the snout 
tip. The first infraorbital overlaps the dorsal half of the maxilla. The posterior 
members of the series cover the cheek region posteriorly to the level of the hyo- 
mandibular. The orbital margin of the infraorbitals is rolled over laterally to provide 
a partial protection for the infraorbital sensory canal. 

Opercular series. The operculum is as wide as it is deep. Its anterior and dorsal 
margins are straight. The posterior margin passes ventrally and slightly posteriorly 
before turning sharply to present an almost horizontal ventral margin. The oper- 
culum forms a comparatively small part of the entire cheek region. 

The suboperculum is a large element as in all albuloids. The posterior and ventral 
margin is smoothly rounded and continues the contour of the operculum. Anteriorly 
the suboperculum sends up a small dorsal wing. The shape of the interoperculum 
could not be determined. 

The preoperculum is incompletely known. The anterior margin lies tightly against 
the hyomandibular, symplectic and quadrate. A deep groove, which is partially 
overlain by bone, runs close to the anterior margin and represents the path taken 
by the preopercular sensory canal. 

Pectoral girdle and Jin. Dorsally the pectoral girdle is attached to the neurocranium 
by the small post-temporal. Three limbs are directed anteriorly from the main 
body of the post-temporal which is represented by a small triangular plate. One 
such projection (the dorsal limb) reaches above the epiotic to which it was no doubt 
connected by ligaments. Another is short, laterally directed and lies in juxtaposition 
to the postero-lateral corner of the pterotic and is homologous with a similar limb 
seen in Albula. The third limb (the ventral limb) passes antero-ventrally to contact 
the neurocranium in the region of the intercalar. 

The supracleithrum overlaps the cleithrum ventrally and is itself overlain dorsally 
by the post-temporal. The supracleithrum is longer than broad, with its anterior 
and posterior margins parallel. From its union with the post-temporal the bone 
slopes postero-ventrally. 

The cleithrum is, as usual, the largest element of the girdle. As seen in lateral 
view the bone is tapered both dorsally and ventrally. The posterior margin slopes 
ventrally at a steep angle and in this respect resembles Pterothrissus. The endo- 
chondral elements are too incompletely known to warrant comment. However, in 
lateral view the posterior limit of these elements lies behind that of the overlying 
cleithrum, as in all albuloids. 

There are at least 12 pectoral fin-rays (Davis 1887 quotes a figure of 14 for all 
species here considered as Lebonichthys). The first ray is unbranched. The 
dichotomy of the inner rays is confined to the distal third of their length. A small 
triangular splint is associated with the base of the outer fin-ray. 

Pelvic girdle and fin. The only evidence of the pelvic girdle is a small, trans- 
versely orientated strip of spongy bone immediately anterior to the bases of the fin- 
rays. The anterior half of the pelvic girdle, if present, was probably represented by 
cartilage. 



FOSSIL AND RECENT 179 

The pelvic fin originates beneath the posterior half of the dorsal fin. The pelvic 
fin is made up of 9-10 rays, the outermost being both the longest and the only 
unbranched member of the series. Associated with the base of the upper half of 
the outer ray there is a large pelvic splint bone. The anterior end of this splint is 
curved dorsally and lies free from the fin-ray bases. 

Vertebral column. One specimen, B.M.N.H. P.9162, is sufficiently well preserved 
for a vertebral count to be made. This specimen has a total of 66 preural vertebrae 
of which approximately 23 are caudal. 

Each centrum is deeper than long. Those of the anterior abdominal region are 
considerably deeper relative to their length than those of the caudal region. The 
centra are marked laterally by four to six longitudinal grooves separated by ridges. 

The neural arches are autogenous throughout the column. In the anterior 
abdominal region the neural arches are expanded distally but in the posterior ab- 
dominal and caudal region such expansions are not seen. The neural spines of the 
anterior abdominal region are fine, straight and the paired halves remain separate. 
Posteriorly the spines of either side are fused to form a median structure which 
exhibits a slight backward curvature. The posterior neural spines are somewhat 
modified to support the caudal fin-rays. 

The parapophyses are autogenous. Throughout most of the abdominal region 
they are represented by small knobs lodged in cup-shaped depressions in the anterior 
half of the centra. The posterior five or six parapophyses are somewhat longer, 
the last being the longest and bearing a correspondingly shorter pleural rib. 

Each pleural rib is fine, slightly curved and incompletely encircles the abdominal 
cavity, substantiating the view that this fish had a rounded body. Except pos- 
teriorly, where the ribs are borne solely by the lengthened parapophyses, the articula- 
tion of the ribs is shared between the parapophysis and the centrum. 

The haemal arches are inserted into depressions in the centra. Distally the arches 
bear slightly curved haemal spines. The posterior five are modified to support the 
caudal rays. 

There are 11-12 supraneurals associated with the first 17-18 neural spines. Each 
supraneural is sigmoid in shape and they form a graded series from large and broad 
anteriorly to small and narrow posteriorly. 

Epineurals are found throughout the column. Anteriorly they are attached to 
the posterior margin of the neural arch/spine junction. In the posterior part of the 
column they are often displaced and in life probably lay free in the myocommata. 

Epipleurals are also present throughout the column. In the abdominal region 
they were closely applied to the parapophyses but throughout the caudal region they 
lay free. The epipleurals in the anterior caudal region show forked bases. 

Median fins. The dorsal fin is relatively long based, the base length being equal 
to 28 per cent of the standard length. The origin of the fin lies near the occiput and 
the occiput-origin distance is considerably less than the distance from the termina- 
tion of the fin to the caudal peduncle, a feature previously noted by Woodward 
(1901 : 16). 

The dorsal fin is composed of 24-25 rays. Davis (1887 : 560, pi. 31, fig. 2) states that 
there are 32 rays but his illustration does not support such a high count. Woodward 



i8o 



ELOPIFORM FISHES 



nsp pu 




Fig. 90. Lebonichthys gracilis (Davis). Caudal skeleton in left lateral view. Arrows 
indicate upper and lower principal fin-rays. Camera lucida drawing of B.M.N.H . P. 4856. 



(1901 : 16) quotes a figure of 20. The first two rays are unsegmeiited. The fifth 
ray is both the longest and the first branched. Of the 21 pterygiophores the first 
three support the first six fin-rays. With the exception of the first three, each 
pterygiophore consists of three distinct radial ossifications. The exceptional 
pterygiophores exhibit fusion between the proximal and middle radials. 

The anal fin arises nearer to the caudal peduncle than to the pelvic fin and consists 
of nine rays supported on seven pterygiophores. The anterior two rays are un- 
segmeiited and the fourth is both the longest and the first branched. 

The caudal fin-rays are supported by five preural and two ural centra. Each of the 
preural centra is deeper than long and marked by ridges. The ural centra are smooth 
and largely overlapped by the uroneurals and the hypural heads. In the specimen 
figured (Text-fig. 90) the second ural centrum is completely obscured by the uro- 
neurals. 

The haemal spines of the five preural vertebrae become progressively stouter from 
the fifth to the first (the parhypural). The parhypural shows the characteristic 



FOSSIL AND RECENT 181 

notch in the posterior margin. This notch lies in juxtaposition to a notch in the 
anterior margin of the first hypural and marks the divergence of two lateral caudal 
arteries from the median dorsal aorta. 

Dorsally, the neural spines associated with the second to fifth preural centra are 
slender, each with a stout posterior edge preceded by a thin median wing. The tips 
of the last three neural spines converge to support the anterior dorsal accessory rays. 
The neural arch complex associated with the first preural centrum appears to be a 
double structure, i.e. two neural arches set one behind the other. This arrangement 
is open to three interpretations. It may represent the two lateral halves of a single 
first preural neural arch, these having become displaced relative to one another 
during preservation (in this connection it is interesting to note that in one specimen 
of Pterothrissus examined the two halves of this arch remain distinct from one 
another except at their extreme tips). Secondly, it may represent a double neural 
arch upon this centrum ; such a condition is quite common in lower teleosts and has 
been recorded for Tarpon (Hollister 1936 : fig. 16). The third interpretation is that 
the posterior structure is the neural arch of the first ural centrum which has become 
displaced anteriorly (cf. Nybelin 1971). Examination of other specimens failed to 
substantiate any of these interpretations. 

There are at least two uroneurals. The first is large and shows a forked base 
overlapping the first and part of the second preural centrum. The second uroneural 
is long, extending from a lateral overlap of the first ural centrum to well beyond the 
distal extremity of the first uroneural (a typical albuloid disposition). Specimen 
B.M.N.H. P. 4856 shows a small element which overlaps the dorsal extremity of the 
second uroneural. Its appearance suggests a uroneural, yet the bone, which shows 
no sign of having been broken, is truncated, unlike a uroneural. Distally this ele- 
ment reaches between the bases of the posterior dorsal basal fulcra. 

The first and second hypurals are supported by the first ural centrum. Distally 
the hypurals support the principal fin-rays of the lower caudal lobe. As is usual 
among lower ' isospondyls ' the first hypural is the larger and has a characteristically 
narrow articular head. In the specimen figured the lower hypurals have been slightly 
displaced postero-ventrally from their true position. 

Four upper hypurals support the principal rays of the upper lobe. The third 
hypural is large and is the only member of the upper series showing a distinct ar- 
ticulatory head. The uppermost hypural is very small and may represent a fusion 
of two elements since there is a faint line of division running down the length of 
this hypural. In the specimen figured the upper hypurals have been displaced 
antero-ventrally. 

Between the first uroneural and the last neural spine there are three epurals, the 
distal extremities of which are obscured by the bases of the basal fulcra which they 
support. 

There are 19 principal rays, 10 in the upper lobe and 9 in the lower. The upper- 
most principal ray is large, unbranched and does not extend over any endochondral 
element (cf. Patterson 1968a). The innermost principal rays do not show expanded 
bases and are thus similar to those of Recent albuloids. The lowermost principal 
ray is supported by the parhypural and the second preural haemal spine. 



ELOPIFORM FISHES 










c 
W 



> 

P 



£ 



s 

o 





fn 




FOSSIL AND RECENT 183 

There are 10- 11 dorsal basal fulcra succeeded by three or four fringing fulcra. 
The posterior five basal fulcra show articulations. Seven basal fulcra occur ventrally 
and are supported by the haemal spines of the third to fifth preural centra. 

An elongate caudal scute occurs above and below the caudal peduncle. 

Squamation. The scales are cycloid. The posterior margin of the scale is rounded, 
the anterior margin straight. Concentrically arranged circuli are apparent in the 
anterior and lateral fields, while the posterior field is marked by a few granulations. 
Anteriorly there are five to eight radii which produce a scalloped anterior margin. 



Lebonichthys lewisi (Davis) 
(Text-fig. 92) 

1887 Clupea lewisii Davis : 571, pi. 33, fig. 1. 

1901 Osmeroides lewisi (Davis) Woodward : 17, pi. 2, fig. 3. 

Diagnosis. Lebonichthys reaching 260 mm S.L. Head equal to 30 per cent S.L. 
Vomer with villiform teeth. Maxilla with small villiform teeth. Dorsal fin situated 
nearer to the caudal peduncle than to the occiput, and composed of 18-19 ra y s - 
Caudal skeleton with two epurals ; the supporting neural spines expanded antero- 
posteriorly ; fringing fulcra absent. Caudal scutes present. Scales without 
scalloped margin. 

Holotype. Nearly complete fish, R.S.M. 1891. 59.91, from the Middle Ceno- 
manian of Hakel, Lebanon. The counterpart is B.M.N.H. P. 4868. 

Material. The holotype, counterpart and B.M.N.H. P.6024 from the Middle 
Cenomanian of Hakel, Lebanon. 

Remarks. Lebonichthys lewisi is known by only two individuals of which one, 
the holotype (part and counterpart), is poorly preserved. Little information on the 
head is available from these specimens. 

Known features of the cranial roof closely resemble those of Lebonichthys gracilis 
and Osmeroides latifrons. The frontals are wide above the otic and orbital regions 
but taper above the slightly elongated snout. A long tubular nasal is present. The 
lateral ethmoid is a large, stout ossification, more robust than that of L. gracilis but 
like that of Albida. There is no direct contact between the lateral ethmoid and para- 
sphenoid as there is in Albula. 

The parasphenoid attains its greatest width beneath the lateral ethmoid, as in 
0. latifrons. Teeth are borne on the ventral surface of the parasphenoid beneath the 
orbit. The parasphenoid teeth are all of the same size but their shape varies from 
conical to hemispherical. The vomerine teeth are villiform ; the limits of the sup- 
porting ossification are not known. 

The hyopalatine series is similar to that of L. gracilis except for two obvious 
differences : the metapterygoid only contacts the symplectic over a short area and 
there are teeth on the ectopterygoid in addition to those on the endopterygoid. The 
presence of ectopterygoid teeth is a primitive feature, known only in L. lewisi among 



I >4 



ELOPIFORM FISHES 

ep1-2 



un2 



npul 



pu 5 



10 mm 




Fig. 92. Lebonichthys lewisi (Davis). Caudal skeleton in left lateral view. Arrows 
indicate upper and lower principal fin-rays. Camera lucida drawing of B.M.N.H. P. 6024. 



albulids. Ectopterygoid teeth of L. lewisi resemble those of 0. latifrons. The 
teeth on the dermopalatine are villiform, as in L. gracilis. 

The upper jaw extends to the level of the middle of the orbit in B.M.N.H. P. 6024 
and to the posterior margin of the orbit in P. 4868. The curvature of the oral surface 
is not as marked as in L. gracilis. The teeth on the premaxilla and maxilla are 
villiform in P.6024, more pointed in P. 4868, but never needle-like as in L. gracilis. 

The mandible extends from the snout to beneath the middle of the orbit in P.6024. 
It is slightly longer in P. 4868. The dentary teeth are set in a band along the oral 
surface and resemble those of the upper jaw ; villiform in P.6024 an d pointed in 
P.4868. 

The operculum is slightly deeper than wide with an oblique ventral margin, thus 
contrasting with that element in L. gracilis. In shape the preoperculum resembles 
that of Osmeroides. The sensory canal appears to have been completely enclosed 
by bone. 

There are at least 15 branchiostegal rays borne by the ceratohyals. When the 
mouth was closed the anterior end of the hyoid bar must have lain between the jaw 
rami. The gular plate is narrower than in L. gracilis. 

The pectoral girdle is not known to differ much from that of L. gracilis. The post- 
temporal of L. lewisi has the dorsal and ventral limbs but a central limb is absent. 



FOSSIL AND RECENT 185 

The cleithrum is large, much of its area being given over to muscle attachment. 
There are at least 10 pectoral fin-rays. The pelvic fin shows 11 rays. 

The vertebral column is composed of approximately 64 vertebrae, of which 24 are 
caudal. Over most of the column the neural arches and spines, haemal arches 
and spines and the parapophyses and ribs are precisely the same as in L. gracilis. 
Differences occur in those preural vertebrae associated with the caudal fin-rays. 
Epineural and epipleural (ventral) intermusculars are present throughout the column. 
There are 22-23 supraneurals, more than in L. gracilis, and this reflects a difference 
in the position of the dorsal fin which in L. lewisi is situated nearer to the caudal 
peduncle than to the occiput. The dorsal fin contains 18 or 19 rays. B.M.N.H. 
P. 4868 shows 10 anal pterygiophores indicating that the anal fin was composed of 
approximately 12 fin-rays ; B.M.N.H. P. 6024, however, has 7 pterygiophores and 

9 ra ys- 

The caudal skeleton of L. lewisi (Text-fig. 92) is more like that of Albula than is 
that of L. gracilis. In L. lewisi the supporting neural and haemal spines are expanded 
antero-posteriorly, there are two epurals and both uroneurals are stout and disposed 
in typical albuloid fashion. The proximal end of the first uroneural is expanded. 
This is similar to L. gracilis but unlike Albula vidpes. As is usual in albulids, the 
neural spine on the second preural centrum is full length. 

What is known about the skull of L. lewisi suggests that it is more primitive than 
that of L. gracilis. The jaws are relatively longer and less specialized, and the 
dentition borne by the dermal jaws, palate and basibranchial elements is also more 
primitive. The caudal skeleton, however, appears more specialized than that of 
L. gracilis, approaching the Albula condition. 



IV. DISCUSSION OF THE ORDER ELOPIFORMES 

(a) Historical 

With the exception of the pterothrissids, the fishes grouped here as the order 
Elopiformes have been referred to the families Elopidae and Albulidae by Woodward 
(1901), Boulenger (1910), Barnard (1925) and Fowler (1936). Pterothrissus (Bathy- 
thrissa), on which the family Pterothrissidae is based, has been associated with the 
Salmonidae by Gunther (1877), and the Clupeidae by Gill (1893). A suggestion by 
Boulenger that Pterothrissus is closely related to Albula was confirmed by Ridewood 
(1904) who placed the genus in the Albulidae. This author also endorsed Wood- 
ward's view (1901) that the Cretaceous Istieus is closely related to Pterothrissus. 
Since that time Istieus and Pterothrissus have been considered as constituting a 
separate family, closely related to the Albulidae (Berg 1940) or as a subfamily of the 
Albulidae (Norman i960) or as members of the family Albulidae (Boulenger 1910). 

The fishes grouped in the Elopidae by Woodward (1901) and others have often 
been divided into two families, the Elopidae and the Megalopidae, which in turn 
were grouped as a higher taxon of equal rank with that containing the Albulidae 
and Pterothrissidae (Berg 1940 ; Jordan 1923). Little indication, other than text 



186 ELOPIFORM FISHES 

order, was given as to the relationship existing between the Elopidae ( + Megalopidae) 
and the Albulidae ( + Pterothrissidae) or to their relationship with other ' clupeiform ' 
fishes. 

Bertin and Arambourg (1958) went one stage further by including the Elopidae 
(containing the megalopids), Albulidae and Pterothrissidae in a suborder Elopoidei 
of the order Clupeiformes (sensu lato). In the Elopoidei Bertin and Arambourg 
{op. cit.) also included the extinct family Thrissopateridae (Thrissopater and Pachy- 
rhizodus), an assemblage thought by the present author to be but distantly related 
to the elopiforms. 

Although the elopids, megalopids, albulids and pterothrissids have been considered 
closely interrelated by the above authors, the reasons put forward are based upon the 
common possession of primitive and retentional features. Characters often quoted 
as indicating relationship are : medially united parietals, gular plate, upper jaw 
formed by both the premaxilla and maxilla, posterior circumorbitals large, opercular 
apparatus complete, scapular foramen entirely enclosed, etc. Ride wood (1904 : 54) 
recognized the fallacy of using such criteria when he stated ' Such resemblances as 
exist between them [albulids and elopids] are explicable by the fact that neither has 
departed to any great extent from the ancestral group from which all the Teleostean 
fishes sprang. . . '. It is indeed true that none of the elopiform fishes has progressed 
far beyond the pholidophorid stage. Elops is justifiably credited with being the 
most archaic of living teleosteans and some authors have gone so far as to suggest that 
Elops is a holostean. Saint-Seine (1949), for instance, places Elops in the Hale- 
costomi and Nybelin (1956) suggested that Elops is as much a holostean as is Amia 
or Lepisosteus. Gosline (1971 : 112, fig. 28) questions a close relationship between 
Elops and Albula, and implies that elopoids and albuloids may each have closer 
relatives outside the Elopiformes. 

Clearly a relationship, if it exists, between the Elopidae, Megalopidae and the 
Albulidae and Pterothrissidae, must be based upon advanced and specialized charac- 
ters inherited from the common ancestor. The presence of rostral ossicles and a 
leptocephalus larva, both considered here to be advanced features, were recognized 
by Bigelow (1963) as evidence of relationship, and this view is supported here. 

The classification of the Elopiformes I have used is essentially that of Greenwood 
et al. (1966) with the addition of a new family, the Osmeroididae, to the Albuloidei. 
I believe that this scheme most closely reflects the phylogeny of the elopiform fishes. 

The history of thought concerning the relationship of the elopiforms to other 
' lower teleosts ' has been stable. Most authors have placed the elopiform families 
near to the clupeids with the implication that the latter are their closest living rela- 
tives (Boulenger 1910 ; Berg 1940 ; Bertin & Arambourg 1958 ; and others). 
Woodward (1901), perhaps significantly, dealt with the osteoglossids and noto- 
pterids between the elopiforms and clupeids. 

Greenwood et al. (1966) consolidated a strengthening opinion that the ' isospondyls ' 
represent an artificial assemblage and proposed a three- or fourfold origin of teleosts 
from pholidophorid ancestors. In this scheme the elopiforms were considered to have 
evolved from the pholidophorids independently of other lower teleostean fishes with 
the possible exception of the Clupeiformes (Clupeidae, Engraulidae, Chirocentridae 



FOSSIL AND RECENT 187 

and Denticipitidae). Subsequent work has indicated that the clupeiforms possibly 
represent yet another attainment of the teleostean grade (Greenwood 1970b). The 
nearest relatives of the Elopiformes were thought by Greenwood et al. (1966) to be the 
Anguilliformes and the Notacanthiformes. Together these orders were grouped in 
a Division I (subsequently named Cohort Taeniopaedia by Greenwood, Myers, 
Rosen & Wietzman 1967) by virtue of the fact that members of all three orders have 
a leptocephalus larva, and certain other characteristics (Greenwood et al. 1966) ; the 
Elopiformes share the presence of rostral ossicles with the Notacanthiformes, and the 
Notacanthiformes are related to the Anguilliformes by a similarity in swimbladder 
anatomy. 

(b) Salient features of the Elopiformes and the relationship with other ' lower teleosts ' 

The following list of salient characters has been drawn up to serve three main 
purposes : to indicate the wealth of primitive characters shown by the elopiforms ; 
to serve as a reference point by which comparisons with other teleosts may be made ; 
and to supplement the diagnosis given on p. 6. 

1. Parietals in contact medially. 

2. Dermethmoid separate from underlying mesethmoid. 

3. Supraoccipital small, supraoccipital crest weakly developed. 

4. Nasals slender, laterally placed. 

5. Neurocranium well ossified in the otic region. 

6. Post-temporal fossa with roof. 

7. Subtemporal fossa large. 

8. Orbitosphenoid and basisphenoid well developed. 

9. Parasphenoid with teeth. 

10. Anterior myodome bone absent. 

11. Pars jugularis long ; separate lateral openings for the hyomandibular ramus 
of VII, jugular vein and orbital artery. 

12. Hyomandibular broad, orientated vertically or nearly so. 

13. Palatine composed of separate autopalatine (or palatine cartilage) and 
dermopalatine. 

14. Premaxilla and maxilla entering gape. Supramaxilla(e) present. 

15. Endochondral elements of the gill arches ossified, remaining separate from 
the dermal elements. 

16. Basibranchial (and usually a basihyal) tooth plate well developed. 

17. Ceratohyal composed of anterior and posterior ossifications, separated by a 
narrow band of cartilage. 

18. Two hypohyals. 

19. Urohyal shallow. 

20. Infraorbital series composed of six elements. The first is large and is followed 
by a narrow second element. 

21. Antorbital with associated sensory canal. 

22. Infraorbital sensory canal branches anteriorly to form an antorbital branch 
and an ' ethmoid ' branch. 



188 ELOPIFORM FISHES 

23. Opercular apparatus complete, with a large suboperculum and large inter- 
operculum. 

24. Pectoral girdle with a mesocoracoid arch articulating ventrally with both the 
scapula and coracoid. 

25. Coracoid large. 

26. Pectoral fin held low down on the body. 

27. Pelvic fin abdominal, held horizontally, with more than 10 rays. 

28. Pelvic splint bone present. 

29. Vertebral centra pierced for the passage of the notochord. Neural arches, 
haemal arches and parapophyses autogenous. 

30. Neural spines in the anterior abdominal region composed of separate lateral 
halves. 

31. Supraneurals forming a complete series between the occiput and the origin 
of the dorsal fin. 

32. Epineural intermuscular bones developed throughout the column. 

33. Caudal fin forked. 

34. Caudal skeleton with two free ural centra, uroneurals free from centra, 
hypurals separate from one another and from the ural centra. Parhypural 
without a hypurapophysis. 

35. Nineteen principal caudal fin-rays ; 10 in the upper lobe, 9 in the lower. 

36. Scales cycloid, with bone cells. 

37. Lateral line complete. 

38. Swimbladder with anterior outpushings developed on either side of the mid- 
line. 

39. Retroarticular absent. 

40. Development of a leptocephalus larva. 

41. Rostral ossicles present. 

42. Pectoral splint present at the base of the outermost fin-ray. 

43. Basipterygoid process absent. 

44. Mesethmoid weakly ossified or absent. 

45. Ectopterygoid process present (its development varies within the order). 

46. Supraorbital and infraorbital canals connected above the sixth infraorbital. 



In addition, there are many other primitive characters found in the basal family 
(the Elopidae) of the order. These are listed below : 

47. Body fusiform, showing little compression, fins acuminate. 

48. Cranium shallow, mouth terminal. 

49. Neurocranium shallow, skull roof flat. 

50. Cephalic sensory canals enclosed by bone. 

51. Supraorbital sensory canal extending to the parietal. 

52. Dilatator fossa with roof. 

53. Ethmoid commissure complete, contained within a bony tube that runs 
across the dermethmoid. 

54. Intercalar large, and with the prootic forming an intercalar-prootic bridge. 



FOSSIL AND RECENT 



189 



55- 

56 
57 

58 

59 
60 
61 

62 

63 
64 
65 

66 
67 
68 

69 
70 



Occipital condyle formed by the exoccipitals and basioccipital. Together 
these elements articulate with a vertebral centrum which is functionally part 
of the neurocranium. 

Vomer short, with a broad head which bears a paired patch of small teeth. 
Anterior ceratohyal with an elongate tooth patch plastered along the dorsal 
margin. 

Branchiostegal rays numerous. 
Gular plate large, horizontal. 

Premaxilla small, forming little of the functional upper jaw. 
Gill arches with suprapharyngobranchials associated with the first and 
second epibranchials. A fifth epibranchial is present. 
Mandible with low coronoid process situated posteriorly. 
Posterior infraorbitals broad, covering much of the cheek. 
Dermosphenotic (sixth infraorbital) large. 

Supratemporals large, meeting one another in the mid-line with the conse- 
quence that the supratemporal commissure is wholly enclosed by bone. 
Anterior uroneural with forked base. 
Seven hypurals. 

Base of the innermost principal fin-ray of each caudal lobe expanded and 
considerably overlapping the supporting hypural. 
Urodermal present. 
Fringing fulcra usually present (but absent in Elops). 



Of the characters listed above, 1-39 and 47-70 are primitive for teleosts, 1 the 
majority being found in the less specialized representatives of the superorders Clupeo- 
morpha, Osteoglossomorpha, Protacanthopterygii (sensu Rosen & Patterson 1969) 
and Ostariophysi. These characters are therefore of little use in indicating rela- 
tionships ; they merely serve to suggest ways in which the elopiforms have not evolved. 
No other group of ' lower ' teleosts shows such a wealth of primitive characters and 
it is clear that the Elopiformes could not have evolved from any extant member (s) 
of the superorders mentioned above. Indeed, as noted below, it is extremely un- 
likely that elopiforms evolved from any known fossil teleost group belonging to the 
other superorders. 



1 Feature 38, the possession of anterior diverticula of the swimbladder is included here as a character 
which is primitive for teleosts. The evidence is admittedly indirect, since there are no details of swim- 
bladder anatomy in fossils except where that organ influences osteological development. Within the 
'lower' teleosts an otophysic connection involving an anterior diverticulum of the swimbladder extend- 
ing forward to the level of the neurocranium has developed in the Megalopidae (Elopomorpha) , the 
Clupeomorpha, and in the Notopteroidei and the young growth stages of Mormyriformes (both Osteo- 
glossomorpha). The development of otophysic connections in these groups, which are presumed to have 
evolved in parallel (Greenwood et al. 1966), suggests that some rudimentary swimbladder modification 
was present in the Pholidophoridae. The stage hypothesized for the pholidophorid condition may be 
seen in Albula, where paired anterior diverticula extend forward to terminate beneath the ventro- 
lateral surfaces of the basioccipital (Greenwood 1970a) ; such diverticula are present, although poorly 
developed, in Elops and Pterothrissus. It is of interest to note that ostariophysan fishes, which are 
thought to have evolved from a group (the Salmoniformes, see Greenwood et al. 1966) without anterior 
diverticula of the swimbladder, developed an otophysic connection by vertebral and rib specializations 
(Rosen & Greenwood, 1970). 



igo ELOPIFORM FISHES 

Features 43-46 are advanced to some degree over the basic teleost condition 
but may be seen in a large number of distantly related teleosts and have probably 
arisen independently in a number of lineages. They are therefore of limited use in 
indicating relationship. 

Only three characters (nos. 40, 41 and 42 2 ) indicate that the Elopiformes represent 
an interrelated assemblage of fishes distinct from other ' lower ' teleosts. Elopiforms 
show the specialized character combination of a leptocephalus, rostral ossicles and 
pectoral splint bone. The leptocephalus larva has been used by Greenwood et al. 
(1966) as a feature relating the Elopiformes, Anguilliformes and Notacanthiformes 
as a distinct, natural lineage. 

A note of caution must be introduced here since it is not certain whether the 
development of the leptocephalus is a specialized condition or if it is simply another 
feature retained from halecostome ancestors. Harrisson (1966) and Romer (1966) 
both imply that the leptocephalus is a retentional feature, but give no evidence for 
this. 

A leptocephalus occurs in three orders (also linked by other intergroup charac- 
teristics, Greenwood et al. 1966) which are anatomically, ecologically and biologically 
diverse. Yet it is present in these three orders only. The absence of a leptocephalus, 
or anything approaching such an ontogenetic stage, in clupeomorphs, osteoglosso- 
morphs or protacanthopterygians is perhaps the strongest argument for assuming 
that this larva evolved only within the elopiforms. Furthermore, the peculiarities 
of the leptocephalus suggest that it arose but once and must have been present in 
early Cretaceous times, before the elopoid-albuloid dichotomy. In short, circum- 
stantial evidence is in favour of the leptocephalus being an elopiform specialization 
subsequently inherited by the eels and notacanths (+ halosaurs and lipogenyids). 

If, in the future, it should be demonstrated that the leptocephalus is a primitive 
or retained feature then its use as an indicator of interordinal relationship will no 
longer be valid and it will have to be added to the already long list of elopiform charac- 
ters that have been inherited from halecostome ancestors. 

A second feature that is here assumed to be an elopiform development is the 
possession of rostral ossicles. The snout of teleosts appears to be very labile, 
subject to many modifications related to feeding mechanisms. The sensory canals 
and associated ossifications are no less labile and may be seen to vary considerably 
both between and within each family. Within the Elopiformes, variation in the 
sensory canals of the snout has involved the movement, loss, or even multiplication 
of rostral ossicles, the existence of which is unique to the Elopiformes and Nota- 
canthiformes. 

The origin of rostral ossicles as seen in Elops has been discussed by Nybelin (1956, 
1967a), who suggested that they arose by fragmentation of the dermethmoid (rostral 

2 (Added in proof.) Jessen (1972) has since published an excellent study of the pectoral girdle and fin 
of actinopterygians. Jessen suggests that the pectoral splint is a fin-ray derivative and is a primitive 
character, being found (in a more primitive state) in Lepisosteus and Amia. Jessen 's conclusions are 
accepted here. The presence of a pectoral splint in elopiforms must therefore be regarded as of little 
use in linking elopoids with albuloids. The pectoral splint is nevertheless diagnostic for elopoids and 
albuloids among teleostean fishes. The present author knows of no other teleost with a pectoral splint, 
implying that neither elopoids nor albuloids could have been derived from known teleosteans. 



FOSSIL AND RECENT 191 

of Nybelin op. cit., producing the lateral rostral ossicle) and the antorbital (producing 
an additional ossicle). Although there may be a prenasal ossicle incorporated in 
the dorsal part of the lateral rostral ossicle (Nybelin 1967a : fig. 2), there is never any 
connection between the supraorbital and ethmoid commissure sensory canals. 

Megalopids differ little from the Elops condition, the most significant difference 
being the loss of the anterior fragment of the antorbital. According to Gosline 
(1961 : 22) the loss of this ossicle has been accompanied by the loss of a small section 
of the canal and thus the ethmoid commissure fails to connect with the infraorbital 
system. Nybelin (1967a), however, finds a connection in the soft tissue of Megalops 
and young Tarpon. Following Nybelin's technique of ink injection I find a connec- 
tion in Megalops and Tarpon (the largest Tarpon examined in this respect was of 
222 mm S.L.). A difference from the Elops condition is the very close relationship 
between the lateral rostral ossicle and the premaxilla. In young Megalops and 
Tarpon the ossicle is clear of the premaxilla (Nybelin 1967a : fig. 2), but in older speci- 
mens of these genera the rostral ossicle comes to lie in a shallow excavation on the 
lateral surface of the premaxilla, and is very tightly bound to that bone. 

While reduction in the number of canal bearing bones may be a feature of the 
megalopids, multiplication is characteristic of the albuloids. Thus Pterothrissus 
has three ossicles, although it is impossible to be sure which are dermethmoid deriva- 
tives and which are antorbital in origin. It is suspected that the fossil albuloids 
had more than three ossicles since the development of the premaxillary canal was 
probably mediated by a morphological stage such as is seen in Recent halosaurs 
(Notacanthif ormes) . 

The development of the snout canal system of albuloids is hypothesized to have 
taken the following course. From a basal albuloid such as Osmeroides (which has 
an ethmoid commissure very much like that of Elops, other ossicles being unknown), 
the anterior end of the dermethmoid turned ventrally through 90 ° taking the rostral 
ossicles with it. The rostral ossicles subsequently fragmented into several smaller 
ossicles, some of which lay in the skin over the premaxilla. Such a stage may be 
seen in the halosaurs Halosaurus and Halosauropsis, where there are four to seven 
ossicles forming a chain across the snout, each ossicle apparently associated with a 
neuromast (Gosline 1961). In albulids and pterothrissids those ossicles overlying 
the premaxilla became closely associated with that bone (as in Tarpon) and eventually 
fused with it, producing the condition seen in the albulids and pterothrissids where 
the sensory canal appears to pass through the premaxilla. The premaxillary canal 
is supplied by a branch of the buccal nerve (indicating its relationship with the 
infraorbital canal) which pierces the premaxilla close to the point of articulation of 
that bone with the ethmoid. The ethmoid commissure may still be continuous in 
some halosaurs (Gosline 1961), but it is broken in the mid-line in Recent albuloids, 
presumably by the loss of a median portion. It appears that in this assemblage 
(albuloids and halosaurs) the ethmoid commissure has shifted its association from 
the dermethmoid to the premaxillae. 

The retention of the originally continuous ethmoid commissure-infraorbital canal 
system in the Elopiformes is probably a consequence of the solid snout of these 
fishes. In other groups, notably clupeiforms, salmoniforms and cypriniforms, the 



192 ELOPIFORM FISHES 

movement of the premaxilla and maxilla is such that a delicate sensory canal would 
be stretched and broken. The isolated ethmoid commissure remains as a superficial 
line of pores within the skin of many ' lower ' teleosts (Nybelin 1967a) . 

The Elopiformes are traditionally linked with the Clupeiformes (Denticipitoidei 
and Clupeoidei) (Regan 1909 ; Berg 1940 ; Bertin & Arambourg 1958). As sug- 
gested by Greenwood et al. (1966) there is little evidence to support this proposal. 
If a relationship exists it is probably to be found below the teleost level. 

Clupeiform anatomy has been influenced by the following trends : 

a. Development of an otoplwsic connection involving an intimate intracranial 
association between the swimbladder and the inner ear. 

b. Development of a recessus lateralis, and elaboration of the sensory canal 
system on the cranial bones. This is often accompanied by the loss of a 
well-defined lateral line canal on the trunk. 

c. Development of a temporal foramen between the frontal and parietal which 
houses a region of sensory tissue, the ' sinus temporalis ' (see Patterson 1970a 
for a discussion of this sensory tissue). 

d. Deepening of the suspensorium with an increasing mobility of the dermal jaw 
elements, particularly the supramaxillae. (The engraulids and chirocentrids are 
exceptional in this respect ; Denticeps has no supramaxillae, Greenwood 1968) . 

e. Development of microphagous feeding habits (except Chirocentrus) which 
principally involves reduction or loss of teeth on the dermal jaws, gill arches 
and palate and the development of epibranchial organs (Nelson 1967). 

f. Development of a unique type of caudal skeleton, characterized by a free 
first hypural, fusion of the second hypural with the first ural centrum and 
fusion of the first uroneural with the first preural centrum (Gosline i960, 
1961 ; Greenwood et al. 1966 ; Ca vender 1966). Denticeps is somewhat more 
primitive (Greenwood 1968). 

g. Lateral compression of the body accompanied by the development of scutes 
in the ventral mid-line (a few also show dorsal scutes, e.g. Diplomystus, 
Hyperlophus and Potamalosa) . 

These clupeiform trends are quite distinct from any seen in the elopiforms. In 
those elopiforms with an otophysic connection (Megalopidae) it is entirely extra- 
cranial and involves hypertrophy of the intercalar, an element which in clupeomorphs 
is reduced to a tiny cap of bone. Lateral compression of the body, also seen in the 
megalopids, is never associated with the development of scutes. 

The Upper Cretaceous species of Diplomystus are in many respects the most primi- 
tive clupeiforms known (Patterson 1967a). A derivation of Diplomystus from a 
basal elopiform, such as the Upper Jurassic Anaethalion vidali, is unlikely in view 
of the fact that Diplomystus has retained a basipterygoid process and a heavily 
ossified mesethmoid (for the primitive and advanced condition of the ethmoid 
elements see Patterson 1970b). Conversely, elopiforms could not have been derived 
from a Diplomystus-like clupeoid since the latter is already specialized towards the 
clupeoid condition in showing pre-epiotic fossae, a specialized path of the supra- 
temporal commissure and certain specializations of the tail (Patterson 1967a : fig. 8 ; 
Greenwood 1968 : fig. 32 ; Cavender 1966 : fig. 4). 



FOSSIL AND RECENT 193 

Thus it is improbable that clupeiforms were derived from known elopiforms or 
vice versa. However, this does not rule out the possibility that these two orders 
shared a common ancestor above the pholidophorid level, within the teleosts. The 
ancestry of clupeiforms is obscure. Prior to the Albian no teleost is known to possess 
any of the primary clupeiform cranial characters (for these, see Patterson 1970a). 
Despite this, both the Clupavidae and the Leptolepididae have been considered 
ancestral to the clupeiforms (Arambourg 1950, 1954). 

Recently, Patterson (1970b) has indicated that the Clupavidae is a heterogeneous 
assemblage containing many poorly known forms. Those species which may be 
referred to the Clupavidae with any degree of certainty show a fusion of the first 
ural centrum with the first preural centrum (Patterson, op. cit.), a feature only found 
among euteleosteans. 

This leaves the leptolepids as candidates for clupeiform ancestors. A suggested 
leptolepid-clupeiform relationship is based upon gross similarities of jaw, trunk and 
fin shape (Arambourg 1950, 1954). At present there is no firm evidence that the 
leptolepids are involved in the ancestry of the clupeiform fishes, yet there is no reason 
to reject such a suggestion. 

The leptolepids could not have been ancestral to the elopiforms, since the latter 
are more primitive in showing a dermethmoid which is separate from the mesethmoid. 
In leptolepids the dermethmoid is fused with the mesethmoid. On the other hand, 
elopiforms are more specialized than leptolepids in possessing a supraorbital-infraor- 
bital sensory canal connection and in lacking a basipterygoid process. There is no 
known teleost primitive enough to be considered ancestral to both the elopiforms and 
the leptolepids or to the elopiforms and the clupeiforms. 

In short, it is very unlikely that the elopiforms and the leptolepids are related 
above the pholidophorid level and it is even more unlikely that the elopiforms and 
the clupeiforms are related within the Teleostei. 

Members of the Elopiformes have at one time or another been considered related 
to members of the Osteoglossomorpha (following the work of Patterson 1967c, the 
plethodonts are included in this assemblage : the ichthyodectids are also considered 
related to osteoglossomorphs, Greenwood et al. 1966 ; Patterson 1967c). It is 
unlikely that such a relationship exists within the Teleostei. 

Garstang (1931) suggested a relationship between the Elopoidei and the Hyo- 
dontoidei + Mormyroidei on the basis that at least some members of all three groups 
show the development of an otophysic connection and a similarity in the opposition 
of parasphenoid and basihyal teeth. However, the type of otophysic connection is 
very different in the two groups and was probably developed independently. The 
similarity in buccal occlusion may be explained as the retention of a primitive 
feature, such as is found in pholidophorids, and is therefore of little use in indicating 
relationship. 

Woodward (1901) related the Osteoglossidae and Plethodontidae to the Albulidae, 
mainly on the basis of the powerful buccal dentition seen in these families. In view 
of the great dissimilarities in the pattern of the roofing bones, infraorbital bones, 
vertebral column and caudal skeleton, the albulids appear far removed from these 
osteoglossomorph fishes (see also p. 207). 

13 



194 ELOPIFORM FISHES 

Bardack (1965) suggests that the plethodonts were derived from elopoids but the 
reasons given do not stand critical examination. The similarity in dentition is a 
convergent development (p. 207), the similarity in neurocranial architecture is merely 
due to primitive characters (flat roof, medially united parietals and roofed post- 
temporal fossae), and the common possession of an ossified interorbital septum is 
again an example of convergence to meet a similar functional need. 

The osteoglossomorphs are anatomically diverse, but the evolutionary trends which 
have influenced osteoglossomorph anatomy are quite distinct from those manifest 
in the elopiform complex. The osteoglossomorphs show : 

a. The development (in the vast majority) of a predatory method of feeding 
which is imposed upon the basic teleostean parasphenoid-basibranchial/basi- 
hyal bite. This trend has resulted in a relatively immobile palate and upper 
jaw and the retention of both a basipterygoid process and a well-ossified 
ethmoid region (for a fuller discussion of osteoglossomorph jaw adaptations 
see Greenwood et al. 1966). 

b. Development of an otophysic connection in some ; but this is never developed 
in either the megalopid or the clupeomorph fashion (Greenwood 1963). 

c. Reduction and consolidation of the circumorbital elements (Nelson 1969b). 

d. Close association between the upper jaw and the infraorbitals. 

e. Development of a unique type of caudal skeleton (Greenwood 1966 ; Cavender 
1966 ; Nelson 1969b) which can only be derived from the type found in Liassic 
leptolepids or pholidophorids. 

The diversity of the Osteoglossomorpha makes overall comparison with the 
elopiforms difficult, but in none of the osteoglossomorph groups are there significant 
features with counterparts in the elopiforms (for Ichthyodectidae see Bardack 1965 ; 
for Tselfatioidei see Patterson 1967c ; for Osteoglossiformes and Mormyriformes see 
Greenwood 1966 and Greenwood et al. 1966). 

Palaeontological evidence suggests that the osteoglossomorph complex is an 
ancient one ; the Hiodontoidei are first known from the Upper Jurassic (Greenwood 
1970b), the highly specialized Tselfatioidei from the Albian, and the Ichthyodectidae 
are recorded from the Tithonian (possibly as early as the Oxfordian, Bardack 

I965)- 

Ichthyodectids such as Allothrissops from the Kimmeridgian, and Thrissops from 
the Oxfordian appear to be the least specialized osteoglossomorph fishes. Even in 
these forms, however, there are features which militate against their having been 
derived from, or having given rise to, elopiform fishes. Allothrissops, for instance, 
is more specialized than elopiforms in showing a reduced supratemporal, a reduction 
in the size of the antorbital and a restricted parasphenoid dentition. It also ap- 
parently shows no distinct dermethmoid element (Patterson 1967a : fig. 2). The 
supraorbital-infraorbital canal connection of Allothrissops is developed in a different 
way from that seen in elopiforms. In Allothrissops the connection is indirect, via 
the otic sensory canal, whereas in elopiforms it is direct. 

Specialized features of primitive elopiforms which rule out their being considered 
ancestral to Allothrissops and related teleosts include : loss of the basipterygoid 
process, reduction in snout ossification, reduction of branching of the preopercular 



FOSSIL AND RECENT 195 

canal, and the more advanced type of caudal skeleton in which there is a differentia- 
tion of the uroneural series (p. 91). 

Between the remainder of the osteoglossomorphs and the elopiforms there is an 
even greater ' incompatibility ' of specialized features, particularly in aspects of 
caudal anatomy. In short, the present state of palaeontological knowledge suggests 
that the osteoglossomorphs and the elopiforms are not related above the pholi- 
dophorid level. 

A comparison between the Elopiformes and primitive members of the Euteleostei 
(salmonoids, osmeroids, esocoids, argentinoids, chanoids and gonorynchoids) is 
difficult since the early history of euteleosteans is very poorly known. That all 
euteleostean fishes were derived from a single teleost ancestor has yet to be demon- 
strated, and the relationships between the primitive euteleosteans mentioned above 
are not clear. It seems certain, however, that the euteleostean complex is an ancient 
one since specialized members occur in Cenomanian times (Goody 1969b ; Patterson 
1967b). In view of the uncertainty concerning the origin and early evolution of the 
euteleosteans, remarks will at this stage be kept brief. 

Elopiformes and each of the euteleostean groups appear divergently specialized 
and any relationship of ancestor-descendant type between these two groups appears 
impossible. However, the possibility of there having been an ancestor common to 
the elopiforms and the euteleosteans cannot be ruled out. 

In this connection the Upper Jurassic Anaethalion is of interest. The anatomy of 
Anaethalion is poorly known, particularly with respect to cranial details. One of 
the contained species, A. vidali from the Kimmeridgian of Spain, shows a pectoral 
splint bone and is thus referred to the Elopidae (p. 36). I have examined a few 
specimens of the species described by Nybelin (1967b) and am unable to determine 
the presence or absence of a comparable splint bone in the Bavarian forms. For 
the time being therefore the systematic position of most Anaethalion species must 
remain in limbo (see also pp. 35-44). I would, however, like to draw attention to 
certain features of the caudal skeleton of the genus. 

The tail of many euteleostean fishes shows several distinctive features. Firstly, 
the last few neural and haemal spines show the development of median expansions 
of laminar bone (Patterson 1970b ; Greenwood & Rosen 1971). Secondly, in some 
euteleosts, there are well-developed neural arches associated with the first preural and 
first ural centra, which in these fishes may fuse with the first uroneural forming the 
stegural (Paterson 1970b). The fate of these neural arches varies during the evo- 
lution of different euteleostean lineages (for a discussion of the primitive euteleostean 
tail see Patterson 1970b). Thirdly, the epural series of euteleosteans is composed of 
two or three epurals which are usually of equal length, unlike the graded series seen 
in the caudal skeleton of elopiforms. 

Within the genus Anaethalion these ' euteleostean ' features are variously deve- 
loped. Two forms mentioned by Nybelin (1967b), A. (?) cf. subovatus and A. sp. 
(' £/o^>s-ahnlicher Fisch ', Nybelin 1963, 1971) show few of the above features. A. 
angustissimus shows laminar bone, moderately well developed neural arches and 
epurals of equal length (Text-fig. 18C, Nybelin 1971). A. knorri (Text-fig. 18D, 
Nybelin, 1971) and A. angustus (Text-fig. 18B) both show laminar bone, but the 



196 ELOPIFORM FISHES 

neural arch associated with the first preural centrum is poorly developed (the neural 
arch of the first ural centrum is not known). The epurals of these two species are 
of unequal length and form a graded series. 

Significantly, the elopiform Anaethalion vidali shows the development of laminar 
bone and moderately expanded neural arches associated with the first preural and 
ural centra, but the epurals form a graded series (Text-fig. 17). 

Anaethalion cannot be regarded as ancestral to both the elopiforms and euteleo- 
steans (assuming the latter to be monophyletic) since there are euteleosts such as 
alepocephaloids and the osmerid-like Humbertia more primitive than Anaethalion 
in possessing a basipterygoid process (for alepocephaloids see Gosline 1969 ; for 
Humbertia see Patterson 1970b). 

Considering the similarity between the caudal skeleton of primitive elopiforms 
and euteleosts, there are two conclusions to be drawn. Either the development of 
laminar bone has little significance in indicating relationship, having been developed 
in parallel, or a true relationship exists, beneath the Anaethalion level, from an an- 
cestor very similar to Anaethalion in which there was a basipterygoid process. 

In summary, the Elopiformes show no relationship with either the Clupeomorpha 
or the Osteoglossomorpha above the pholidophorid level. A relationship with the 
euteleosts is suggested by the caudal anatomy of Anaethalion but definite evidence is 
lacking. 

The order Elopiformes is constituted by two radically different suborders, the 
Elopoidei and the Albuloidei. Each suborder shows trends not seen in the other, 
and it is difficult to discern any trend common to both. The elopoids retain the 
greatest number of primitive characters and appear in the fossil record well before 
the albuloids. On these criteria (and see below) the albuloids are considered to have 
evolved from the elopoids. 

Albuloids may be traced back to the Albian, where they are represented by 
Osmeroides lewesiensis, a form that exhibits an essentially albuloid neurocranium 
but resembles elopids rather than other albuloids in many primitive features (e.g. 
48, 49, 50, 51, 55, 56, 58, 59, 60, 62, 63, 68 and possibly 69 from the list on p. 189). 
The dichotomy between the albuloids and the elopoids must therefore have taken 
place at some time prior to the mid-Albian. Within the elopoids only the primitive 
Elopidae could have been ancestral to the albuloids. The other elopoid family, the 
Megalopidae, shows trends very different from those seen in albuloids, such as deepen- 
ing of the head and body, enlarged scales, enlarged post-temporal fossae and the 
development of an otophysic connection. 

(c) Suborder Elopoidei 

The suborder Elopoidei consists of the conservative Elopidae and the more 
specialized Megalopidae. Opinion is divided as to whether all elopoids should be 
grouped in a single family or whether to recognize two families, as is done here. 
Many authors, writing from a primarily neontological standpoint, recognize a single 
family, the Elopidae (Boulenger 1910 ; Barnard 1925 ; Fowler 1936 ; Bertin & 
Arambourg 1958 ; Okada i960 ; Hildebrand 1963). Hildebrand (1963) speaks 
against separate family status for Elops on the one hand and Megalops (and Tarpon) 



FOSSIL AND RECENT 197 

on the other, quoting similarities in primitive characters (gular plate, numerous 
branchiostegals, large maxillaries, etc.) to support his argument. Others have 
adopted the families Elopidae and Megalopidae (e.g. Jordan 1923 ; Berg 1940 ; 
Danil'chenko 1964 ; Romer 1966). It is perhaps significant that these latter authors 
considered both fossil and Recent forms. Fishes grouped here as megalopids have 
a long history, extending back at least to the Albian. Furthermore, megalopids 
show several specializations of long standing that would not be appreciated by 
including the megalopids with the elopids. Therefore, in recognition of the long 
separate history and the fact that the megalopids show several specializations not 
seen in the Elopidae, the megalopids are accorded separate familial status. 

The Elopidae, the most primitive of the Elopiformes, are represented by Elops, 
Davichthys and at least one species of Anaethalion (A. vidali). The primitive nature 
of Elops has been stressed by many authors and both Saint-Seine (1949) and Nybelin 
(1956) have suggested that Elops is a halecostome. This view has little to support it. 
Firstly, Elops shows basic teleostean features such as perichordally ossified centra, 
fin-rays of the lower caudal lobe that are supported by two hypurals which articulate 
with a single centrum, modification of the ural neural arches to form uroneurals 
(these caudal features have been used by Patterson 1967a, 1968a to define the 
Teleostei), the development of a complete lateral commissure, and narial openings 
which are situated close together and distinct from the nasal bone. Secondly, to 
include Elops in the Halecostomi would result in all elopiforms, notacanthiforms and 
anguilliforms also being considered as halecostomes and the term teleost would, in 
consequence, have little meaning. 

The Elopidae are recognized as a group distinct from other elopiforms by the 
absence of specializations. This is clearly an unsatisfactory situation but one that 
is very difficult to rectify. Trends within the fishes grouped here are few and minor : 
the snout becomes slightly longer, with a resulting increase in the length of the 
dermethmoid ; the quadrate/mandibular articulation shifts slightly posteriorly 
with a corresponding slight increase in mandibular length ; the first infraorbital 
develops an ascending limb which reaches the supraorbital and thus excludes the 
antorbital from the orbital margin ; the preopercular sensory canal moves nearer 
to the anterior margin of the bone and there is a reduction in branching of the infra- 
orbital and preopercular sensory canals. A reduction in the branching of the 
sensory canals is seen in all elopiforms and the upward growth of the first infra- 
orbital is seen in megalopids. 

Elops itself cannot be ancestral to any other elopiform since it lacks fringing fulcra 
(present in megalopids and early albulids) and the ventro-lateral projections asso- 
ciated with the dermethmoid (present in megalopids). Davichthys, from the 
Cenomanian and Upper Santonian, retains these primitive features, but occurs too 
late in time to be considered ancestral to either the megalopids or the albuloids, since 
both the latter groups must have originated prior to the Albian. Interest thus 
centres on Anaethalion vidali, an Upper Jurassic elopid, which does show certain 
features normally associated with the megalopids (see p. 42). 

The Megalopidae contains six (possibly seven) genera which show distinct speciali- 
zations. The differences between the megalopids and their closest relatives, the 



198 



ELOPIFORM FISHES 



elopids, are given below (Table I). 
specialized. 



In all these features the megalopids are more 



Table I 





Differei 


ices between the Elop 


idae and the Megalopidae 




Character 


Elopidae 




Megalopidae 


i. 


Body 


Rounded 




Compressed 


ii. 


Neurocranium 


Shallow 




Deepened 


iii. 


Quadrate/mandibular 
articulation 


Behind level of eye 




Beneath eye 


iv. 


Middle pit-line 


Present 




Absent 


v. 


Autosphenotic spine 


Weakly developed 




Well-developed 


vi. 


Epiotic process 


Weakly developed 




Well-developed 


vii. 


Post-temporal fossae 


Reaching anteriorly 


to the 


Reaching anteriorly to the 






level of the autosphenotic ; 


level of the orbitosphenoid ; 






never confluent 




confluent in those that can be 
examined in this respect 


viii. 


Intercalar 


Moderately large 




Large and of complex shape 


ix. 


Otophysic connection 


Absent 




Present 


X. 


Rostral ossicles 


Two 




One, representing the more anterior 
of elopids 


xi. 


Pelvic fins 


Originating beneath 


or 


Originating beneath or anterior 






posterior to the dorsal fin 


to the dorsal fin 


xii. 


Anal fin 


Short based, first 
pterygiophore short 




Long based, first pterygiophore 
long 


xiii. 


Scales 


Lateral line tubes 




Lateral line tubes branched 



unbranched 



Trends within the Megalopidae include : the enlargement of the post-temporal 
fossae with a corresponding increase in the convexity of the skull roof ; reduction of 
parasphenoid teeth ; the development of a superior mouth and the development 
of a large coronoid process on the mandible. 

The megalopid neurocranium has been modified in response to the forward exten- 
sion of the swimbladder and epaxial musculature. In primitive teleosts the bulk 
of the epaxial musculature extends forwards into paired post-temporal fossae whose 
anterior wall is formed by the autosphenotic of either side. In Recent megalopids 
these fossae have become enlarged so that the anterior wall of the fossa is formed by 
the orbitosphenoid and the fossae of either side are confluent above the cranial vault. 
The vertical depth of these fossae has also increased resulting in greater convexity 
of the skull roof in the otic region. Many teleosts have retained post-temporal 
fossae, but the enlargement of these vacuities is a feature unique to the megalopids. 
Some other groups, e.g. salmonids and ichthyodectids, have supplemented the post- 
temporal fossae by the development of grooves upon the roof, while a general ten- 
dency among clupeomorphs and more advanced protacanthopterygians is to erode 
the roof so producing a post-temporal groove. 

The otophysic connection of the Megalopidae is of a simple type, totally unlike 
that found in other groups possessing a swimbladder-cranial linkage (clupeomorphs, 
notopteroids and juvenile mormyriforms, Greenwood et at. 1966). Although the 



FOSSIL AND RECENT 199 

megalopid otophysic connection was probably derived from a basic type (see foot- 
note, p. 189) common to the other groups mentioned, its subsequent development 
has followed a different and restricted path. There is never any direct connection 
between the swimbladder and the endocranial cavity and the association of the 
precoelomic diverticulum with the intercalar is unique. 

There are several minor osteological and soft anatomical differences between the 
otophysic connection of Megalops and Tarpon (Greenwood 1970a), but there is 
nothing to suggest that both types were not derived from a common ancestral type 
very much like that of Tarpon. The Eocene megalopids exhibit a simple type of 
otophysic connection in which the lateral cranial wall has a single large depression 
for the swollen end of the swimbladder, which presumably lay against the skull at 
this point as in Tarpon. The diverticulum could not have been as closely associated 
with the cranium as it is in the Recent megalopids since the vagus foramen, the 
glossopharyngeal foramen and the posterior opening of the jugular canal still occupy 
primitive positions. In the Recent megalopids the nerve foramina have moved to 
allow the swimbladder to fit tightly under the intercalar. Thus in both Tarpon and 
Megalops the glossopharyngeal leaves the neurocranium immediately beneath the 
vagus and slightly behind the level of the intercalar. The posterior opening of the 
jugular canal has moved upwards, so much so in Megalops that it opens directly 
into the subtemporal fossa, and the head vein has acquired a different spatial rela- 
tionship to the intercalar (see p. 67). 

The intercalar of the Recent megalopids has been described in detail by Greenwood 
(1970a) and shown to be more complex in Megalops than in Tarpon. However, in 
both there is a lower intercalar limb associated with the swimbladder diverticulum. 
Specimens of the Eocene megalopids do not show a lower intercalar limb yet it is 
assumed that one was present and that it was developed much as in young Tarpon 
(see Greenwood 1970a) . If the lower intercalar limb was absent in the Eocene species 
one would be forced to recognize that an identical intercalar development had taken 
place twice, a view with little to recommend it. A lower intercalar limb must have 
been present in Protarpon as well as in Promegalops since the latter has a neurocranium 
more like that of the more specialized Megalops than Tarpon. The apparent absence 
of the lower intercalar limb in the fossil representatives is not surprising in view of 
the fragility of the pedicel connecting the lower limb to the main body of the inter- 
calar. 

The caudal anatomy of the Megalopidae is primitive and essentially similar to the 
Elops-type. Loss of the urodermal is derived relative to the Elops condition, but 
the retention of fringing fulcra is primitive. 

Fringing fulcra are rare among teleosts. Besides the Megalopidae, fringing fulcra 
are known in Leptolepis (Leptolepididae), Lebonichthys (Albulidae), Anaethalion 
(? Elopidae) and Davichthys (Elopidae). All members of the Megalopidae in which 
the caudal fin is known show fringing fulcra. Fringing fulcra in teleosts are con- 
fined to the region above the uppermost principal caudal raj', and, except in Seden- 
horstia and older specimens of Megalops, only one or two are usually present. 

In the teleosts mentioned above, the upper principal caudal ray is preceded by a 
number of fringing fulcra in front of which lie a number of fin-rays variously termed 



200 ELOPIFORM FISHES 

epaxial fin-rays (Gardiner 1970), raylets (Hollister 1936; this term applied to those 
rays which are not articulated), accessory rays (Goody 1969b), procurrent rays 
(Patterson 1968b) or basal fulcra (Patterson 1968a). Irrespective of terminology, 
these rays, which are supported by epurals or neural spines, appear to represent 
fringing fulcra that have migrated downwards (Gardiner 1970). In support of this 
interpretation is the fact that, in most lineages, fringing fulcra occur prior to epaxial 
rays, the exception being the pholidopleurid Australosomus. 

It is hypothesized that in the primitive actinopterygian tail there were no fin-rays 
above the axis of the notochord. All rays now found in this position are derived 
from fringing fulcra that have moved down and forwards. The downward move- 
ment of fringing fulcra has occurred in the lineages leading to the pholidopleurids, 
pycnodonts, perleidids, pachycormids, amioids and teleosts, and in all instances it is 
associated with tail shortening. Having moved down, the fringing fulcra become 
basal fulcra (epaxial fin-rays, etc....), supported by endochondral elements. 
Subsequent elongation of the basal fulcra produces the need for articulation. Mega- 
lops and Tarpon are unusual among teleosts retaining fringing fulcra in having 
articulated basal fulcra, although a great many teleosts without fringing fulcra have 
articulated basal fulcra. 

A more vexatious problem concerns the origin of the fringing fulcra themselves. 
Reasonable suggestions as to their origin fall into three categories : firstly, derivation 
from the median scale row ; secondly, derivation from the repeated unilateral 
dichotomy of the leading fin-ray ; thirdly, breaking up of the articulated leading ray. 
For the first hypothesis there is little evidence. A median scale row would form an 
equally effective cutwater as a fulcral element formed of two lateral halves, and 
there is no reason to believe that scales broke into lateral halves. 

The second hypothesis, that fringing fulcra arose by unilateral dichotomy of the 
leading fin-ray, was suggested by Gardiner (1970). Such a development followed 
loss of the cutwater previously provided by the median scale row and the scaled 
body lobe of the tail. That fringing fulcra are not found in early actinopterygians 
with unbranched fin-rays supports this view. Furthermore, the first occurrence of 
caudal fringing fulcra is concurrent with the first occurrence of caudal ray branching 
(Stegotrachelus finlayi, upper Middle Devonian of Scotland) . However, if fringing 
fulcra were formed by branching of the leading ray, numerical correspondence would 
be expected between the fulcra and the segments of the supporting ray. Ideally, 
one fulcrum would be associated with one articulation. This is certainly not the 
case in the Elonichthyidae (Gardiner 1970), and in a range of genera examined there 
was never any constant numerical relationship between the fulcra and the articula- 
tions of the supporting ray. One final point is that the number of fulcra would be 
fixed at an early ontogenetic stage, when, although not necessarily ossified, the rays 
exhibit the adult branched condition (observations on young Tarpon). 

In Megalops (and to a lesser extent in Tarpon) the number of caudal fringing fulcra 
increases throughout the life of the individual (Text-fig. 34), long after the caudal 
rays have ossified. Furthermore, the ray immediately anterior to the uppermost 
principal in Megalops shows articulations which lie at an increasingly oblique angle 
distally ; it is possible that the posterior articulations become separated as fringing 



FOSSIL AND RECENT 201 

fulcra. This is the third hypothesis mentioned above. It is interesting to note 
that in some pholidophorids (Pholidophorus bechei and Pholidolepis dorsetensis, see 
Patterson 1968a) there is a reduced ray which lies at the base of the fringing fulcral 
series and which perhaps represents the basal articulation of a ray which has frag- 
mented distally into many fringing fulcra. 

A distinctive feature of both Megalops and Tarpon is the elongate last ray of the 
dorsal fin. This feature is of little use in indicating relationships since it has 
developed in Dorosoma and Opisthonema (Clupeidae) and also in the albulid Dixonina. 
In Tarpon and Megalops the growth of this terminal filament shows positive allo- 
metry. The independent development of the terminal filament in the three groups 
of fishes mentioned may have functional significance. The dorsal fin-rays of Mega- 
lops and Tarpon are relatively immobile, with a weak basal musculature. Most of 
the rays are oval in cross-section but in the last ray the filament is ' T '-shaped 
with the stem of the ' T ' directed anteriorly. The wings of the ' T ' are broad, 
particularly at the base of the ray, and extend laterally from the plane of the fin. 
When erect, this filament will produce turbulence resulting in drag opposing the 
forward momentum of the fish. I suggest that the flexible filament is forced to a 
nearly horizontal position when the fish is swimming rapidly, with the water pressure 
acting against the erector muscles. In this horizontal position the filament would 
produce minimum drag. As the speed of the fish decreases the filament will become 
more vertical, producing a drag component which is located behind the centre of 
gravity. In this position the drag would be useful in aiding the fish to decelerate 
before the flexure of the body and caudal fin allow the animal to turn. 

Lineages within the Megalopidae. Sedenhorstia stands apart from other mega- 
lopids in both primitive and advanced characters. The primitive characters include 
the small scales, simple dorsal fin without an elongated terminal filament, and 
shallow lower jaw without a prominent coronoid process. Specialized features 
include the dentition, which is represented by a single row of teeth in the upper and 
lower jaws, the ossifications within the dorsal ligament, and fusion of the neural 
arches associated with the first preural and first ural centra. Such specializations, 
although few and minor, divorce Sedenhorstia from the ancestry of Eocene mega- 
lopids. Furthermore, Sedenhorstia occurs too late (Cenomanian-Campanian) to be 
ancestral to a more ' typical ' megalopid such as Elopoides. 

The Albian Elopoides exhibits no specializations that would preclude it from con- 
sideration as an ancestor to the later megalopids. The depth of the cranium and 
the steep profile of the skull roof suggest that the post-temporal fossae were already 
large in this form, and the megalopid nature of the dentary, maxilla and scales is 
evident. Mention was made above (p. 85) of the apparent resemblance of Elopoides 
to Megalops rather than Tarpon, particularly with respect to the depth of the cranium 
and the large orbit. These features may be interrelated as a function of the absolute 
size of the fish. A small fish may be expected to have a relatively large eye, and 
hence a greater bulk of eye musculature accommodated within a deepened myodome. 
This, in turn, would effectively reduce the depth of the branchial chamber. To 
compensate, the suspensorium became deeper. Thus the apparent resemblance to 
Megalops may not have phyletic significance in this particular case. In other 



202 ELOPIFORM FISHES 

features Elopoides may be considered ancestral to either or both of the Eocene 
megalopids. 

Of the Eocene megalopids, Protarpon and Promegalops, the former is more primitive 
and resembles Tarpon. Promegalops is closely related to Megalops, the similarities 
including the shape of the neurocranium and the position of the lower jaw articula- 
tion. Whether Promegalops is more closely related to Protarpon or Elopoides is 
unknown, there being no evidence one way or the other. 

(d) Suborder Albuloidei 

The suborder Albuloidei consists of three families, the Albulidae, Pterothrissidae 
and Osmeroididae (nov.). The suborder is known in the fossil record from the Albian 
and is represented in the Recent fauna by three genera (Albula, Dixonina and Ptero- 
thrissus). The Albulidae and Pterothrissidae are very much alike, more so than 
either is to the stem family, the Osmeroididae, and in the discussion of the families 
the first two are considered together. The Osmeroididae show few of the specializa- 
tions of the Albulidae or the Pterothrissidae. 

Fishes grouped here as the Albulidae and Pterothrissidae have been recognized 
as a distinct assemblage but placed near the elopoid fishes because of the common 
possession of many primitive features (Ridewood 1904). The Osmeroididae, 
containing the genus Osmeroides and possibly Dinelops, have in the past been asso- 
ciated with the Elopidae (Woodward 1901 and all subsequent authors). Although 
Osmeroides does not show many albuloid specializations, it shares with these fishes 
a pattern of neurocranial architecture which is significantly more advanced than 
that of the Elopidae and completely different from that of the Megalopidae. Fur- 
thermore, within the genus Osmeroides can be seen the initiation of morphological 
trends that were to become fully expressed in both the Albulidae and the Ptero- 
thrissidae. In other words, the Osmeroididae are a link between the Elopidae on 
the one hand and the Albulidae and Pterothrissidae on the other. The Osmeroididae 
are included in the Albuloidei on the basis of neurocranial specializations. 

In all albuloids the neurocranium is shallow and the subtemporal fossa is deep, 
extending inwards to partially occlude the post-temporal fossa. The latter is small 
and directed antero-medially. The sub-epiotic fossae augment the post-temporal 
fossae in receiving the epaxial trunk musculature. The dilatator fossae are always 
deep, narrow dorso-ventrally and completely roofed. In the lateral neurocranial 
wall there is a large ridge running across the face of the prootic which provides a 
point of attachment for the anterior branchial musculature. The otic bulla, con- 
taining the sacculith, is very large and precludes any possibility of there being a 
primitive type of otophysic connection (cf. Megalopidae). There is no prootic- 
intercalar bridge, indeed the intercalar is reduced in albuloids. The lateral ethmoid 
and parasphenoid meet, sometimes in a suture. Finally, there is a deep depression 
at the base of the parasphenoid ascending wing, the significance of which is not known. 

Other advanced cranial characters found in nearly all albuloids are the enlarged 
sensory canals (developed to varying degrees throughout the group), a reduced 
supratemporal, resulting in the middle portion of the supratemporal commissure 
running in the skin, and the position of the quadrate/mandibular articulation which 



FOSSIL AND RECENT 203 

is always beneath the orbit or the lateral ethmoid. The trunk of albuloids is rounded 
and the head is nearly as broad at the occiput as it is deep. 

Trends and characters seen in the more 'advanced ' albuloids include the following : 
i. Enlargement and modification of the sensory canals, 
ii. Roof of the neurocranium with anterior ridges and troughs, particularly in 

the snout region, which is slightly elongated, 
iii. Hyopalatine series long, shallow, with an anteriorly directed quadrate. 
The symplectic is large and the palatine has processes articulating with both 
the mesethmoid and the lateral ethmoid, 
iv. Vomer reduced. 

v. Mouth small, inferior and specialized for bottom feeding. 
vi. Premaxilla forming much of the oral margin of the upper jaw, maxilla 
reduced, with its head inturned and moving independently of the ' fixed ' 
premaxilla. A single supramaxilla. 
vii. Differentiation and modification of the palatal and basibranchial dentition, 
viii. Specialization of the dentition associated with the gill arches. 
ix. Dentary with a strongly developed, inflected ventral border, and correlated 
with this a reduction in size of the gular plate which may (Albula) be 
vertical. 
x. Dermal jaw dentition of fine needle-like teeth. 

xi. Modification of the caudal skeleton, involving reduction in the number of 
epurals, hypurals and uroneurals. 
The albuloid neurocranium is distinctive in showing a slightly protruding snout 
which, in dorsal view, is narrow and marked by ridges and troughs associated with 
the enlarged supraorbital sensory canal. The Osmeroididae generally have no promi- 
nent ridges on the roof but Osmeroides latifrons shows a neurocranial roof much like 
later albuloids. Enlargement of the sensory canals is seen throughout the group. 
In the basal Osmeroides lewesiensis the cephalic sensory canals run entirely within 
the bones, opening to the surface by pores. In 0. latifrons the supraorbital canal 
opens to the surface above the lateral ethmoid and continues forward in a shallow 
trough, which probably also contained the nasals. In this respect this form resembles 
the Albulidae. In the Pterothrissidae the supraorbital canal runs within bone for 
a short distance only ; thus in Pterothrissus the canal opens above the posterior 
half of the orbit. The otic canal of the Pterothrissidae and of Lebonichthys among 
the Albulidae is only partially covered by bone. 

In the Albulidae, and to a greater extent in the Pterothrissidae, the infraorbital, 
mandibular and preopercular canals are only partially covered by bone. The open 
infraorbital canal results from the coalescence of adjacent sensory pores, a trend 
which is first seen in 0. latifrons. The enlargement and opening out of the man- 
dibular canal has also taken place within Osmeroides. Thus 0. lewesiensis and 0. 
levis show an intramural canal, while the only portion of the mandibular canal 
contained within bone in 0. latifrons is in the articular. In both the Albulidae and 
the Pterothrissidae the mandibular canal lies in a shallow groove, never passing 
through bone. Associated with enlargement of the mandibular canal, the ventral 
edge of the dentary becomes inturned so that in the Albulidae and Pterothrissidae 



20 4 ELOPIFORM FISHES 

the lower margins of the rami meet in the ventral mid-line. The skin between the 
jaw rami is folded in the Recent albuloids and in Albula the gular plate, which is 
necessarily reduced in size, lies vertically. The Pterothrissidae have lost the gular plate. 

The sensory canal system of the snout in the Albulidae and Pterothrissidae is 
peculiarly modified and provides a possible link with the Notacanthiformes. Primi- 
tively, as in 0. lewesiensis, the infraorbital canal ran from the first infraorbital to the 
dermethmoid where it joined its partner of the opposite side through the ethmoid 
commissure. Although they have not been found, it is assumed that there were 
canal-bearing rostral ossicles between the dermethmoid and the first infraorbital in 
0. lewesiensis. In the Albulidae and Pterothrissidae the snout has turned downwards 
and the ethmoid commissure, no longer in its primitive position, is not a continuous 
transverse canal. Rather the infraorbital canal now descends through rostral 
ossicles on to the premaxilla. The extension of the infraorbital canal on to the pre- 
maxilla is a development unique to albuloids and notacanthiforms. 

The premaxilla of 0. latifrons is unknown but it is suspected that canals were 
developed on the premaxilla in this form among the Osmeroididae. 

In addition to the modified infraorbital canal, in the Pterothrissidae the supra- 
orbital sensory canals of either side unite across the dermethmoid. In Albula the 
supraorbital sensory canals are separated anteriorly by a median cavity which is 
isolated from the canals on either side by membrane (Gosline 1961). The similarity 
of this condition to that in anguilliform fishes has been noted by Gosline (1961). 

In the majority of ' lower ' teleostean fishes the sensory canals open to the surface 
by large pores. In the Albulidae and Pterothrissidae, however, the canals are 
covered by taut skin perforated by many tiny pores. The canals themselves are 
probably filled with water and the taut skin acts like a tympanum, resonating in 
sympathy with impinging vibrations. With the exception of the supraorbital, otic 
and temporal canals, the sensory canals are directed ventro-laterally and ventrally 
and accord with the benthic habit of the albuloids. 

There are many other fishes with a cavernous canal system. The majority are 
inhabitants of deep water (e.g. Macrouridae), but there are also shallow-water forms 
(e.g. Notopteridae). The presence of a cavernous sensory canal system is therefore 
not necessarily related to habitat, and while its significance (other than perhaps 
increasing sensitivity) is not understood, its advanced nature is undoubted. 

With the slightly elongated snout and the short, anteriorly situated lower jaw, the 
hyopalatine apparatus of the Albulidae and Pterothrissidae has become modified. 
The hyopalatine series is long and shallow with the articular face of the quadrate 
directed anteriorly. The shifting of the quadrate/mandibular articulation has been 
achieved by forward displacement of the quadrate which is linked to the hyo- 
mandibular by a somewhat lengthened and flattened symplectic. The metaptery- 
goid has come to support the symplectic. A condition very much like that seen in 
Recent albuloids is found in 0. latifrons (Text-fig. 59). 

A further development in the hyopalatine series of most albuloids is the large 
foramen between the hyomandibular and the metapterygoid, through which pass 
deeper fibres of the levator arcus palatini muscle. In 0. lewesiensis and 0. latifrons, 
as in the elopoids and some other teleosts, the metapterygoid bears a ridge on the 



FOSSIL AND RECENT 205 

medial side of which the levator arcus palatini inserts. The albuloids possessing 
the above mentioned foramen do not show such a ridge. The only other ' lower ' 
teleosts I have examined in which there is a comparable foramen are Chaetoessus 
and Chanos, but in these forms there is apparently no passage of musculature 
through this aperture. 

The ectopterygoid of albulid and pterothrissid fishes is produced dorsally into a 
well-defined process which lies against, and supports, the flimsy infraorbital bones 
beneath the orbit. A comparable structure is found in most elopiforms but it never 
reaches the same proportions as in the albuloids. In 0. latifrons this ectopterygoid 
process, although a rudimentary spine-like structure, is set in typical albuloid 
fashion. Elsewhere an ectopterygoid process is rare but does occur in the clupeid 
Dussumieria (Ridewood 1904). 

The palatine is, as in all elopiforms, composed of distinct endoskeletal and dermal 
components. The former is often cartilaginous (ossified in 0. lewesiensis) and 
articulates with the neurocranium at two points, anteriorly with the mesethmoid or 
ethmoid cartilage and posteriorly with the lateral ethmoid. Among elopiforms, 
the lateral ethmoid projection of the palatine is only found in albuloids. Such a 
palatine-neurocranial connection gives some rigidity to the palate. Elsewhere, a 
close and rigid juxtaposition of palatine and lateral ethmoid is found in predatory 
fishes taking large prey such as the Chirocentridae and, presumably, the fossil 
Ichthyodectidae, and is a development in response to a strong biting force. 

The upper and lower jaws of the Albulidae and Pterothrissidae are distinctive and 
produce an inferior mouth suitable for bottom feeding. Osmeroides shows a terminal 
mouth (the condition in 0. latifrons is not known) while in Lebonichthys and Istieus 
it is only slightly inferior. The inferior mouth is produced by a combination of 
snout elongation and shortening of the lower jaw. Jaw teeth of later albuloids are 
found only on the premaxilla and dentary (a few maxillary teeth remain in Istieus 
and Pterothrissus) and the reduced maxilla supports a single supramaxilla. 

The premaxilla is firmly bound to the ethmoid region by ligaments and is capable 
of only the slightest lateral movement. The head of the maxilla is turned sharply 
inwards and in the Albulidae and Pterothrissidae there is no separate palatine process 
on the maxilla as there is in the Osmeroididae and the Elopoidei. Instead, the 
rounded palatine head fits against the rear of the maxillary head. The maxilla of 
Recent albuloids is capable of limited independent movement ; only when the mouth 
is fully open does it enter the gape. 

The dentition of the Albulidae and Pterothrissidae is significantly different from 
the primitive type seen in 0. lewesiensis. In albulids and pterothrissids the dentition 
borne by the maxilla and the premaxilla consists of a broad band of needle-like teeth 
in contrast to the small villiform teeth borne by these elements in 0. lewesiensis. 
The shape of the teeth of 0. latifrons is intermediate. 

The palatal dentition is different in each of the three albuloid families. The 
Osmeroididae are primitive in having small, villiform teeth on the parasphenoid, 
dermopalatine, endopterygoid, ectopterygoid, basihyal and basibranchial. The 
teeth of 0. lewesiensis and 0. levis are more or less uniform in size but those of 0. 
latifrons differ both in size and shape. The albulid dentition is modified for crushing 



206 ELOPIFORM FISHES 

shelled invertebrates and the parasphenoid is equipped with large hemispherical 
teeth which are opposed by similar teeth on the basibranchial plate. 

The pterothrissids have slightly enlarged, conical teeth on the parasphenoid, 
endopterygoid, basibranchial and basihyal tooth plates. There are generally fewer 
teeth than in other albuloids ; indeed in Pterothrissus the parasphenoid teeth, although 
well developed, are confined to the region beneath the parasphenoid ascending wings. 
The difference in dentition no doubt reflects a difference in diet between the families, 
the albulids feeding essentially on molluscs and crabs while the Recent pterothrissids 
prey on the thinner shelled crustaceans. Stomach contents of Pterothrissus gissu 
taken off the coast of Japan (depth unknown) revealed an almost exclusive diet of 
caridean malacostracans. 

Crushing dentition, of which one example is seen in the Albulidae, is of relatively 
rare but repeated occurrence among actinopterygians. The condition is clearly 
specialized and suited to fishes of a restricted ecological niche. Within the Chon- 
drostei, such fishes as the Amphicentridae and the Bobasatraniidae have a crushing 
dentition principally involving occlusion between the pterygoid and coronoid teeth. 
Bobasatrania has a basibranchial plate with crushing teeth (Nelson 1969a). 

In the Holostei the Pycnodontiformes (and to a lesser degree the Semionotidae) 
show a crushing dentition developed upon the dentary and the so called ' splenials ' 
(held to be the result of fusion between the articular, coronoid and splenial, Nursall 
1964). Unlike the chondrostean examples, the opposing dentition in pycnodonts 
is borne by the enlarged vomer, which together with the parasphenoid is keyed to 
the stout ethmoid. These fishes generally have short jaws. 

Among teleosts a crushing dentition is more often developed in various ways 
within the buccal cavity and pharynx. In Albulidae, Phyllodontidae and Pletho- 
dontidae the grinding dentition is upon the basibranchium and parasphenoid (some- 
times also the endopterygoid). Such a development is the result of specialization 
involving the basic or primary teleostean bite. Pristolepis is unique among acantho- 
pterygians in showing a toothed parasphenoid opposed by a crushing dentition on the 
basibranchial complex. However, here a secondarily enlarged basihyal is involved 
(Nelson 1969a). 

Other teleosts exhibiting a grinding or crushing dentition show specialization of 
the posterior gill arch elements. Thus, some cyprinids have molariform teeth on the 
fifth ceratobranchials which work against a callus pad on the ventral surface of the 
neurocranium. The labroids have opposed upper and lower pharyngeal plates. A 
development of grinding teeth in the posterior part of the branchial arches is con- 
comitant upon the possession of a protrusile jaw mechanism and mobile pharyn- 
geals, the latter correlated with the development of retractores arcuum branchialium 
muscles. All these features are found in fishes which have lost the primary teleostean 
bite. 

A crushing dentition has therefore arisen several times and in each case the teeth 
involved have been determined by the ' evolutionary grade ' ot the fishes concerned. 

Apart from the Phyllodontidae, about which very little is known, some of the 
plethodonts have a dentition somewhat similar to the Albulidae. This similarity 
has led to a suggested relationship between these two groups. Woodward (1901) 



FOSSIL AND RECENT 207 

placed the plethodont Ananogmius with the albulids and expressed doubt as to 
whether Plethodus, Thryptodus and Pseudothryptodus should be affiliated with the 
Osteoglossidae or the Albulidae. There are a great many characters suggesting 
that the plethodonts are in no way related to the albulids, including details of the 
cranial roof, infraorbitals, pectoral girdle, vertebral column and caudal skeleton. 
With respect to the dentition, although the teeth of plethodonts are developed on 
the same bones as in albulids, those that show a definite adaptation towards crushing 
(Plethodus and Ananogmius) have molariform teeth coalesced into enlarged dental 
plates with a histology resembling that of the ' tubular dentine ' of dipnoans (Pat- 
terson 1967c). This is in contrast to the well-defined, discrete teeth of albulids. 
Together with the many other anatomical differences, especially the trenchant differ- 
ences in caudal anatomy, this suggests that the albulids are not related to the pletho- 
donts and that the crushing dentition in these groups is a convergent development. 

The dentition on the parasphenoid of albulids needs support to resist the strong 
upward bite. Thus it is not surprising to find an ossified interorbital septum in 
Albula and Dixonina. The albulid interorbital ossification is formed by ventral 
extension of the orbitosphenoid and the enlarged basisphenoid. The parasphenoid 
is further braced by sutural contact with the lateral ethmoid. A similar method of 
bracing the parasphenoid is seen in 0. lewesiensis, which has an equally large para- 
sphenoid. In this form, however, the interorbital septum is incomplete. It is 
interesting to note that in some plethodonts (Patterson 1967c : 227) the interorbital 
septum is also ossified. The sporadic occurrence of an ossified interorbital septum 
suggests that this character is not a good phylogenetic criterion. 0. latifrons, 
which has a very broad parasphenoid, and Lebonichthys, in which the parasphenoid 
is virtually identical with that of Albula, both lack an ossified interorbital septum. 

The branchial arches of albuloid fishes (only completely known in the Recent 
forms) differ from the primitive elopoid type in lacking a fifth epibranchial and in 
having the first suprapharyngobranchial represented by cartilage only (the second 
is absent in albuloids). The dentition associated with the arches is more distinctive. 
In the Albulidae and Pterothrissidae the basibranchial tooth plate carries a specialized 
dentition (see above). The dentition on the other gill arch elements consists of a 
few rounded tooth plates which are clothed with tiny, needle-like teeth. These 
tooth plates are few in number and totally unlike the elopoid condition, where the 
tooth plates form a continuous coating on the oral surface of the gill arches. The 
albuloid gill-rakers are few in number and only developed to any extent on the first 
branchial arch. Each gill-raker is club-shaped, the stem being narrow while the head 
bears a cluster of minute teeth similar to those on the tooth plates. 

The dentition associated with the branchial arches of the Osmeroididae is incom- 
pletely known. The basibranchial tooth plate of 0. levis is very much like that of 
Elops and is thus primitive. Upper pharyngeal plates of 0. latifrons are rounded, 
as in albulids and pterothrissids. There is no substantial guide in the known gill 
arch anatomy to suggest a link between the Osmeroididae and other albuloids. 

Among elopiforms the albuloids show a distinctive hyoid bar, with the upper and 
lower hypohyals offset, both with respect to one another and to the anterior cerato- 
hyal. The afferent mandibular artery (which in most lower teleosts runs through 



208 ELOPIFORM FISHES 

both hypohyals) runs through the lower hypohyal only, before piercing the cerato- 
hyal to emerge on the dorsal surface of that bone at the bottom of a prominent groove. 
The afferent mandibular artery of Albula is contained within the anterior ceratohyal 
and never passes through a hypohyal. 

The paired fins, girdles, median fins and preural vertebral column of albuloids are 
very similar to those of other Elopiformes, retaining such characters as pelvic splint 
bones, autogenous neural and haemal arches and parapophyses, and both epipleural 
and epineural intermuscular bones. In keeping with the rounded trunk the pleural 
ribs tend to be short, incompletely encircling the body cavity. 

Apart from the primitive caudal skeleton of 0. lewesiensis (which differs from Elops 
only in possibly having a fourth uroneural) the albuloid tail is distinctive. There 
are only two uroneurals, the first of which is large and extends to the first preural 
centrum (in Istieus it just covers the second preural centrum). The second uroneural 
extends considerably beyond the distal tip of the first while in the Pterothrissidae 
the proximal end of the second uroneural lies well above the second ural centrum. 
There are six hypurals, one less than in elopoids. The bases of the inner fin-rays of 
each caudal lobe are unexpanded, unlike many primitive teleosts, including elopoids. 
There are many basal fulcra above and below the principal rays, and fringing fulcra 
are retained in Lebonichthys. There are minor details in which the caudal skeleton 
of the albulids differs from the pterothrissids and these are mentioned in the discussion 
of those families. 

The similarity in neurocranial architecture between the Osmeroididae on the one 
hand and the Albulidae and Pterothrissidae on the other is held to be of paramount 
importance and has led to the inclusion of the Osmeroididae within the Albuloidei. 
The more advanced osmeroidids show trends in other cranial bones which are 
decidedly albuloid. 

The differences between the osmeroidids and other albuloids are essentially those 
in which the former family is primitive. These primitive features (with respect to 
the other albuloids) include : a terminal mouth ; the simple maxilla, with a distinct 
palatine head and forming a substantial portion of the functional jaw ; the presence 
of two supramaxillae ; the dentition ; the sensory canal system, which is never 
cavernous although it may be slightly enlarged ; the ornamented cranial bones ; and 
the caudal skeleton (at least in 0. lewesiensis) which has more than two uroneurals 
and in which the bases of the inner rays are expanded. 

Osmeroides is the type osmeroidid genus and contains species known from the Albian 
to Coniacian of Europe and possibly Japan (Yabe & Okada 1930). 0. lewesiensis 
appears to be the most primitive species, with the closely related 0. levis and the more 
distant 0. latifrons as derivatives. Of the three species, 0. latifrons appears closest 
to the ancestry of both the Albulidae and Pterothrissidae. 

The great similarity between the Albulidae and Pterothrissidae suggests that they 
were derived from a common ancestor. Important similarities have been men- 
tioned above but may be summarized as the following : snout elongated, mouth 
inferior ; premaxilla with fine, needle-like teeth bearing a sensory canal and forming 
most of the upper jaw margin ; hyopalatine series with a long palatine, an ectoptery- 
goid process and a hyomandibular-metapterygoid foramen ; branchial arches with 



FOSSIL AND RECENT 



209 



a few club-shaped gill-rakers, dental plates with minute needle-like teeth ; cavernous 
sensory canal system ; caudal skeleton with two uroneurals and six hypurals. 

The origin of these two families is to be found in the Osmeroididae. The Albulidae 
and Pterothrissidae differ from one another in several respects (some of which have 
been referred to above) . Each family has retained and developed features not found 
in the other and these are listed in Table II. 



Table II 
Differences between the Albulidae and the Pterothrissidae 



A Ibulidae 

i. Gular plate present 
ii. No median connection of supraorbital 

canals 
iii. Parietals shorter than broad and 

irregular in shape 
iv. Interorbital septum often ossified 
v. Parasphenoid broad, sutured with lateral 

ethmoid 
vi. Dentition of grinding teeth on 

parasphenoid and endopterygoid, 

opposed by basibranchial tooth plate 

bearing a similar dentition (basihyal 

tooth plate edentulous or absent) 
vii. Maxillary teeth absent (except in 

Lebonichthys lewisi) 
/iii. Vomer and palatine with needle-like 

teeth 
ix. Infraorbitals behind eye completely 

covering the cheek region 
x. Branchiostegals not less than 12 in 

number 
xi. Dorsal fin short (slightly elongate in 

Lebonichthys), anal short 
xii. Caudal skeleton with complete neural 

spine on second preural centrum, caudal 

scute present above and below peduncle. 

Hypurals broad, narrow gap between the 

second and third hypurals. Usually two 

epurals 



Pterothrissidae 

Gular plate absent 

Supraorbital canals of either side united 

Parietals square or longer than broad and 

regular in shape 

Interorbital septum never ossified 

Parasphenoid narrow, not sutured with 

lateral ethmoid 

Dentition of conical teeth on 

parasphenoid, endopterygoid and 

ectopterygoid opposed by basihyal and 

basibranchial tooth plates bearing a 

similar dentition 

Maxillary teeth present 

Vomer and palatine edentulous 

Infraorbitals behind eye narrow and 
incompletely covering the cheek region 
Never more than 10 branchiostegals 

Dorsal fin long, extending along most of 
back, anal slightly elongated 
Caudal skeleton with a half spine on 
second preural centrum, no caudal 
scutes. Hypurals slender, large gap 
between second and third hypurals. 
Usually three epurals 



Many of the above features are self-explanatory and little discussion is given here. 

Some of the differences are specializations due to varying diets (features iv, v, vi) 
or different body form (xi). The reduction of the cheek covering (ix) in the 
Pterothrissidae may be related to the reputed deep-sea habit since a reduction of 
ossification is often noted in fishes from this environment. In features i and ii the 
Pterothrissidae are more advanced while the Albulidae are more advanced in features 
iii and vii. 



M 



210 ELOPIFORM FISHES 

Since there is a trend to reduce the dentition, the albulids may be held to be more 
primitive in the retention of palatine and vomerine teeth (feature viii). However, 
the teeth borne by these elements are needle-like and differ from the teeth in a similar 
position in the Osmeroididae. The vomerine teeth in albulids are raised on a short, 
transversely orientated pedicel to bring them on a level with the palatine teeth. 
Together these teeth form a complete semicircle of backwardly pointed teeth lying 
within a similarly shaped band of premaxillary teeth. Thus, while their retention 
may be primitive, their form is specialized. 

The coronoid process of albulids is situated posteriorly while that in Pterothrissus 
among pterothrissids is situated anteriorly and is forwardly directed. This differ- 
ence may be related to the position of the quadrate/mandibular articulation, which 
lies beneath the posterior half of the orbit in the Pterothrissidae and the anterior half 
of the orbit in the Albulidae. The anteriorly situated coronoid process in Ptero- 
thrissus has the same disposition with respect to the lateral ethmoid as in Albula, 
where the process is situated posteriorly. The insertion of the A 2 + A 3 division 
of the adductor mandibulae is correlated with the position of the coronoid process 
and it is suggested that in Pterothrissus the forwardly situated coronoid process has 
in some measure compensated for the more posterior position of the jaw articulation. 

The difference between the families in the number of branchiostegal rays (feature 
x) may also be related to the position of the quadrate/mandibular articulation. In 
both Albula and Pterothrissus the jaw rami are close together in the ventral mid-line 
and the anterior end of the hyoid bar (Tchernavin 1953) does not extend between the 
jaws to any great extent. The distance between the ventral end of the hyomandibular 
and the rear end of the mandible is relatively greater in Albula than in Pterothrissus 
and the hyoid bar is longer. As the number of branchiostegal rays is proportional 
to the length of the attachment area (Gosline 1967) the pterothrissids would be 
expected to show a lower branchiostegal count. 

It is of interest to note that in Lebonichthys (Albulidae) the large gular plate 
separates the jaw rami and although the jaw articulation is in the ' pterothrissid 
position ' the hyoid bar is long and extends well between the jaw rami. The loss, 
or great reduction, of the gular plate may have indirectly affected the shortening of 
the hyoid bar. 

The Albulidae are known by representatives from the Eocene of Africa and 
Europe and the Upper Cretaceous of Asia and North America. All but one of these 
fossil representatives have been found in shallow-water deposits, which is in agree- 
ment with the habitat of the Recent Albula vulpes. The exception is the Albula from 
the Campanian of Alabama (Mooreville Chalk) which is reported to be a deep water 
deposit (Applegate 1970). The evolution of the Albulidae probably began in Albian 
or Lower Cenomanian times since Lebonichthys lewisi (Middle Cenomanian) already 
shows an albulid tail and well-developed hemispherical teeth within the mouth. 

The Pterothrissidae are poorly represented in the fossil record, the only well known 
fish being Istieus from the Campanian of Germany. It is possible that Hajulia 
from the Middle Cenomanian of Lebanon is also a pterothrissid (see p. 152). In 
part this lack of a fossil record may be a consequence of the deep-water habit, which 
was probably adopted early in their history. 



FOSSIL AND RECENT 211 

(e) Possible albuloid derivatives 

A suggestion that the Notacanthiformes were derived from early Elopiformes was 
put forward by Greenwood et al. (1966 : 355). This suggestion was based upon the 
common possession of a leptocephalus larva and the presence of a bone-enclosed 
ethmoid commissure. Those authors further suggested that the halosaurid ethmoid 
commissure agreed with that of albulid fishes, stressing the presence in both of an 
enlarged canal system in this region. The development of such a canal system is 
dealt with in the general discussion of the elopomorph snout (p. 190). All that need 
be said here is that the development of the albuloid-notacanth snout canal system 
is unique, peculiar, and provides good evidence of a common origin. 

Aside from the canal system there are two other specializations shared by albuloids 
and notacanths : the snout is slightly elongate and fragile, and the small mouth is 
inferior ; the maxilla has a slender inturned head and, with the premaxilla, is capable 
of lateral movement but very little vertical movement. 

Beyond this, the notacanths are very specialized in comparison with the albuloids. 
Some of the specializations of notacanths are probably correlated with their deep-sea 
habitat. Thus the cranial bones are exceedingly thin and several are lost (auto- 
palatine, dermosphenotic, epiotic, basisphenoid, orbitosphenoid, supraorbital, 
lateral ethmoid, pterosphenoid and, in the Notacanthidae, the interhyal). Other 
elements are reduced (interhyal, vomer). 

The trend towards an anguilliform body has resulted in a very reduced caudal 
skeleton and a freeing of the pectoral girdle (at least in Notacanthidae) from the 
cranium. These and other specializations are dealt with at length by McDowell (in 
press). All that is intended here is to convey an idea of the extreme divergence of 
the notacanthiforms. 

Fossil notacanthiforms are rare (in part this may be a result of their deep-sea 
habit) but halosaurs are known from the Campanian (Echidnocephalus) and are 
already so specialized that their halosaurid affinities are undoubted. Thus the origin 
of the group must have been some time previously, and the features shared with the 
Albulidae and the Pterothrissidae suggest that a Cenomanian age would be a 
reasonable estimate. The pterothrissids seem the best candidates for the ancestors 
of the notacanths since this group retains maxillary teeth (found in halosaurs) and 
the palatal dentition is not particularly specialized. The albulid dentition on the 
other hand is already specialized by Cenomanian times. 

Cenomanian pterothrissids are rare, being questionably represented by Hajulia. 
This, together with the apparent absence in the fossil record of earlier, less specialized 
yet identifiable notacanthiforms, obscures the origin of the ' Heteromi '. Neverthe- 
less, a close association with the albuloids (in particular the Pterothrissidae) is strongly 
suspected. 

(f ) Comments on the family Phyllodontidae 

The Phyllodontidae Dartevelle & Casier (1943) have been referred to the suborder 
Albuloidei by Estes (1969a) on the basis of similarity to Albula in the basibranchial 
tooth plate (in particular, the basibranchial scars) and the occlusion pattern with the 
parasphenoid. 



212 ELOPIFORM FISHES 

The Phyllodontidae are related amongst themselves and distinguished from the 
Albulidae by having a phyllodont pattern of tooth replacement and in lacking ptery- 
goid tooth plates (Estes 1969a : 328). Estes considers the absence of pterygoid 
tooth plates to be the more important of these characters. 

The Albian genus Casierius Estes (1969b) has been placed in the Albulidae even 
though tooth replacement in this form is phyllodont. Estes (1969b) uses a similarity 
in the contours of the basibranchial plate and the presence of associated pterygoid 
tooth plates as justification for the inclusion of Casierius within the Albulidae. 

It appears to the present author that Casierius, in showing a phyllodont dentition, 
is clearly related to the Phyllodontidae. To dissociate it from the phyllodonts only 
because it has pterygoid tooth plates (Estes 1969b) seems untenable. The poten- 
tiality to replace teeth in phyllodont fashion, either from directly beneath (Phyllo- 
dontinae) or in alternate fashion (Paralbulinae), would appear a trenchant difference 
from the albulids. 

The first albulid appears in the Lower Cenomanian while Casierius is found in 
the Albian, and by this time it already had a wide geographical distribution in- 
dicating an earlier origin. The phyllodont dentition is clearly more specialized 
than the albulid type and, disregarding the time factor, a derivation of phyllodontids 
from albulids would appear more logical than the converse. 

If albulids were present in the Lower Cretaceous it would be expected that tooth 
plates referable to this family should have been discovered, since these are as robust 
as those of the phyllodontids. One is left with three possibilities regarding the 
position of the phyllodontids : they were derived from the Osmeroididae separately 
from the Albulidae and at an earlier time ; they stand ancestral to the Albulidae ; 
or they evolved in parallel to the Albulidae. 

It is difficult to comment on the possibility that the phyllodontids were derived 
from the Osmeroididae. Certainly no known osmeroidid shows a phyllodont denti- 
tion or any kind of replacement pattern comparable with the Phyllodontidae. 

That the Phyllodontidae stand ancestral to the Albulidae is unlikely since the 
phyllodont dentition is so advantageous in grinding food that to revert to a much 
simpler type of dentition yet still retain a similar feeding method is unlikely. 

A separate but parallel evolution of the Albulidae and the Phyllodontidae appears 
possible. Phyllodontids and Casierius are only known by tooth plates which cannot 
be referred to any more completely known fish. Until phyllodontid material other 
than tooth plates is known it is better to accept the classification within the phyllo- 
dontids as being artificial. The fact that the tooth plates belong to a fish at about 
the ' elopiform grade ' seems a reasonable supposition but the possibility of holostean 
or halecostome affinities cannot be ruled out. 

V. SUMMARY 

Descriptions are given of the Recent and fossil Elopiformes represented in the 
collections of the British Museum (Natural History). The basic systematic conclu- 
sions are as follows : 

1. The Elopiformes are a primitive group of teleostean fishes distinguished 
from other teleostean orders by the combination of three primary characters : a 



FOSSIL AND RECENT 213 

leptocephalus larva, the development of rostral ossicles and the modification of the 
outermost pectoral fin-ray to form a pectoral splint. The view expressed by Green- 
wood et al. (1966) that the presence of a leptocephalus larva links the Elopiformes 
with the Anguilliformes and Notacanthiformes is accepted here. Much work remains 
to be done on anguilliform relationships, particularly with respect to their being 
considered elopiform derivatives. Of these three orders the Elopiformes is the most 
primitive and represents the basal group from which the other two were derived. 

2. Elopiformes are represented in the fossil record and Recent fauna by between 
fifteen and twenty genera. The fossil record extends back to the Upper Jurassic 
where the earliest member is Anaethalion vidali, a species which shows certain body 
proportions seen in the megalopid elopiforms. 

The genus Anaethalion is held to be of considerable interest. In the present state 
of knowledge only one species, A . vidali, may be referred to the Elopiformes since it 
shows a pectoral splint, other species being unknown in this respect. The caudal 
skeletons of some species show a distinct resemblance to Elops while others show 
certain euteleostean characteristics. It may be of significance that the caudal 
skeleton of Anaethalion vidali shows laminar bone, as in euteleostean fishes. In- 
sufficient information concerning the anatomy of the various species referred to 
Anaethalion renders a taxonomic revision premature. For this reason these species 
are referred to as the ' Anaethalion generic complex '. 

3. Within the Elopiformes the Elopidae is the most primitive family and is 
represented by Davichthys gen. nov., Elops and probably some species of Anaethalion. 
The elopids have remained virtually unchanged since the Upper Jurassic. The 
only trends noted are reduction in the branching of the cephalic sensory canals, 
loss of the ventro-lateral projections associated with the dermethmoid and loss of 
fringing fulcra. 

4. The Megalopidae, represented by Megalops, Tarpon, Promegalops, Protarpon 
gen. nov., Elopoides, Sedenhorstia and possibly Pachythrissops, were an early deriva- 
tive of the Elopidae. The chief trends characterizing this family are : the develop- 
ment of a superior mouth ; an increase in size of the post-temporal fossae ; the 
development of a unique type of otophysic connection ; reduction of parasphenoid 
teeth ; and certain minor modifications in the postcranial skeleton. 

5. The suborder Albuloidei is known from the Albian to Recent. The most 
primitive albuloids are represented by the genus Osmeroides on which the new family 
Osmeroididae is based. Osmeroides is a link between the Elopidae on the one hand 
and the Albulidae and Pterothrissidae on the other. Within the genus Osmeroides 
there is seen the initiation of morphological trends that were to become fully ex- 
pressed in the Albulidae and Pterothrissidae. Osmeroides lewesiensis is the most 
primitive species, showing an external morphology very much like that of the elopids. 
The neurocranium, however, is albuloid in such features as small antero-medially 
directed post-temporal fossae, sub-epiotic fossae, the absence of a prootic-intercalar 
bridge, a small intercalar and a deep dilatator fossa. Osmeroides latifrons is a more 
advanced species and exhibits a slightly elongated snout, opening out of the 
cephalic sensory canals, specialization of the dentition within the mouth and a 
short lower jaw with a strongly inflected ventral margin. 



2i 4 ELOPIFORM FISHES 

6. The Albulidae and Pterothrissidae are similar to one another in many advanced 
features such as : the pattern of sensory canals ; the inferior mouth in which the 
immobile premaxilla is the major functional component of the upper jaw ; the 
development of a hyopalatine series characterized by a hyomandibular-metaptery- 
goid foramen, a flattened symplectic, a well-developed ectopterygoid process and a 
double articulation of the palatine with the neurocranium ; certain specializations 
of the dentition borne by the gill arches ; and loss of the seventh hypural and third 
uroneural in the caudal skeleton. The shared specializations indicate a common 
ancestry for the two families which is to be found within the Osmeroididae. 

The Albulidae and the Pterothrissidae are divergently specialized in features 
related to different diets. Both families may be traced back to Cenomanian times. 

7. The Notacanthiformes show certain similarities in snout morphology with 
the albuloids indicating that the albuloids may stand ancestral to that order. The 
Recent albuloids (the only albuloids sufficiently well known in details of snout 
morphology) are too specialized to be considered ancestral, but the possibility of some 
Cretaceous albuloid ancestry of notacanthiforms is suspected, the early pterothrissids 
being the most likely candidates. 

8. Relationships of Elopiformes with basal groups of other cohorts are briefly 
reviewed. It is suggested that there is no relationship within the Teleostei between 
elopiforms and the Clupeomorpha or between the elopiforms and the Osteoglosso- 
morpha. The possibility of a relationship between the elopiforms and any euteleo- 
stean group is difficult to analyse because of the sparse information available on the 
early evolution of the more primitive members of the Euteleostei. Attention is 
drawn to a few minor features of caudal anatomy seen in both euteleosts and 
elopiforms which may indicate the existence of a common teleostean ancestor. 



VI. ACKNOWLEDGEMENTS 

I wish to thank Dr H. W. Ball, Keeper of Palaeontology in the British Museum 
(Natural History), for the opportunity of studying the collections which form the 
basis of this work. I am grateful to Dr B. G. Gardiner of Queen Elizabeth College 
(London University) for his supervision of the work ; to Professor G. Chapman in 
whose Department the work was done ; to Drs C. Patterson and P. H. Greenwood 
of the British Museum (Natural History) for fruitful discussion, and to the authorities 
of Queen Elizabeth College for the provision of a Research Demonstratorship grant. 



VII. REFERENCES 

Agassiz, L. 1833-1844. Recherches sur les Poissons fossiles : 5 vols., 1420 pp., 396 pis., with 

supplement. Neuchatel. (For dates of publication see Woodward and Sherborn, 1890.) 
1845. Rapport sur les poissons fossiles de l'argile de Londres. Rep. Br. Ass. Advmt Sci. 

York, 14 (1844) : 279-310. 
Allis, E. P. 1903. The skull and the cranial and first spinal nerves in Scomber scomber. J. 

Morpk., Philadelphia, 18 : 45-328, 10 pis. 
Applegate, S. P. 1970. The vertebrate fauna of the Selma Formation of Alabama. Fieldiana, 

Geol. Mem., Chicago, 3 : 383-433, 30 figs. 



FOSSIL AND RECENT 215 

Arambourg, C. 1950. Nouvelles observations sur les Halecostomes et l'origine des Clupeidae. 

C. r. hebd. Seanc Acad. Sci., Paris, 231 : 416-418, 1 fig. 
1952. Les Vertebres fossiles des gisements de Phosphates. (Maroc, Algerie, Tunisie.) 

Notes M6m. Serv. Mines Carte, geol. Maroc, Rabat, 92 : 1-372, 44 pis. 

1954- Les poissons cretaces du Jebel Tselfat. Notes Mem. Serv. Mines Carte, geol. Maroc, 

Rabat, 118 : 1-188, 20 pis., 68 figs. 

Bardack, D. 1965. Anatomy and evolution of chirocentrid fishes. Paleont. Contr. Univ. 

Kansas, Lawrence, Vertebrata 10 : 1-88, 2 pis., 27 figs. 
Barnard, K. H. 1925. Marine Fishes of South Africa. Ann. S. Afr. Mus., Cape Town, 

21 : vii + 418 pp., 17 pis., 17 figs, 
de Beaufort, L. F. 1909. Die Schwimmblase der Malacopterygier. Morph. Jb., Leipzig, 

39 : 526-644, 1 pi., 10 figs. 
Berg, L. S. 1940. Classification of fishes both Recent and fossil (first edition, Russian with 

complete English translation). Trudy Zool. Inst., Leningr. 5, 2 : 517 pp., 190 figs. 
Bertin, L. & Arambourg, C. 1958. Superordre des Teleosteens : 2204-2500, figs. 1561-1788. 

In Traite de Zoologie, 13, ed. Grasse, P. -P. Masson et Cie, Paris. 
Bigelow, H. B. 1963. Suborder Elopoidea. In Fishes of the Western North Atlantic. 

Mem. Sears Fdn mar. Res., New Haven, 1, 3 : xxi + 630 pp., 132 figs. 
Bleeker, P. 1859. Enumeratio specierum fiscium hucusque in Archipelago indico observa- 

tarum. Act. Soc. Sci. Indo-Neerl., Batavia, 6: 1-276. 
Bonaparte, C. L. J. L. 1846. Osservazoni sullo stato della Zoologica in Europei in quanto 

ai Vertebrati nell'anno : 1-51, 8 pis. Firenze. 
Boulenger, G. A. 1910. Teleostei. In The Cambridge Natural History, 7, ed. Harmer and 

Shipley : xvii + 760 pp., 440 figs. Macmillan & Co., London. 
Casier, E. 1946. La faune ichthyologique de l'Ypresien de la Belgique. Mim. Mus. r. 

Hist. nat. Belg., Bruxelles, 104 : 1-267, ° pis., 18 figs. 

1966. Faune ichthyologique du London Clay. Text + Atlas : i-xiv + 496 pp., 68 pis., 

82 figs. Br. Mus. (Nat. Hist.), London. 

Ca vender, T. 1966. The caudal skeleton of the Cretaceous teleosts Xiphactinus, Ichthyodectes, 

and Gillicus, and its bearing on their relationship with Chirocentrus. Occ. Pap. Mus. Zool. 

Univ. Mich., Ann Arbor, 650 : 1-15, 1 pi. 
Cobban, W. A. & Reeside, J. B. 1962. Correlation of the Cretaceous formations of the western 

interior of the United States. Bull. geol. Soc. Am., New York, 63 : 1011-1044, 1 pi., 2 figs. 
Cockerell, T. D. A. 1912. Observations on fish scales. Bull. Bur. Fish., Wash., 32: 117-174, 

52 figs., 8 pis. 

1933- The antiquity of Albula. Copeia, Washington : 226. 

Cope, E. D. 1872. On the families of fishes of the Cretaceous formation in Kansas. Proc. 
Am. phil. Soc, Philadelphia, 12 : 327-357. 

1877. On some new or little known reptiles and fishes of the Cretaceous number 3 of 

Kansas. Proc. Am. phil. Soc, Philadelphia, 17 : 176-181. 

Cuvier, G. & Valenciennes, M. A. 1846. Histoire naturelle des Poissons, 19 : 398 pp., Paris. 
Danil'chenko, P. T. 1964. Superorder Teleostei : 396-472, pis. 5-14. In Osnovy Palaeon- 

tologii, 11, Beschelyustnye ryby. ed. Obruchev, D. B. Akad. Nauk SSSR., Moskva. (In 

Russian.) 
Dartevelle, E. & Casier, E. 1943. Les poissons fossiles du Bas-Congo et des regions voisines. 

Ann. Mus. r. Congo beige, Tervuren, A (3), 2, 1 : 200 pp., 16 pis. 
David, L. R. 1940. Upper Cretaceous fish remains from the western border of the San 

Joaquin Valley, California. Pubis Carnegie Instn, Washington, 551 : 83-112, 3 pis., 

11 figs. 
Davis, J. W. 1887. The fossil fishes of the Chalk of Mount Lebanon in Syria. Scient. Trans. 

R. Dubl. Soc, 3, (2) : 457-636, 25 pis. 
Derschied, J. M. 1924. Structure de l'organe olfactif chez les poissons. Premiere partie. 

Osteichthyes, Teleostei, Malacopterygii. Annls Soc. r. zool. Belg., Bruxelles, 54: 79-162, 

26 figs. 



216 ELOPIFORM FISHES 

Dixon, F. 1850. The geology and fossils of the Tertiary and Cretaceous Formations of Sussex : 
422 pp., 40 pis. F. Dixon, London. 

Dunkle, D. H. 1940. The cranial osteology of Notelops brama (Agassiz), an elopid fish from 
the Cretaceous of Brazil. Lloydia, Menasha, 3 : 157-190, 9 figs. 

Eastman, C. R. 1914. Catalog of the fossil fishes in the Carnegie Museum. Part 4. Descrip- 
tive catalog of the fossil fishes from the lithographic stone of Solenhofen, Bavaria. Mem. 
Carneg. Mus., Pittsburg, 6 : 389-423, 17 pis. 

d'ERASMO, G. 1946. L'ittiofauna cretacea dei dintorni di Comeno nel Carso Triestino. Atti 
Accad. Set. fis. mat. Napoli, (3a) 2, 8 : 136 pp., 1 pi., 34 figs. 

Estes, R. 1969a. Studies on fossil phyllodont fishes : Interrelationships and evolution in 
the Phyllodontidae (Albuloidei). Copeia, Washington : 317-331, 8 figs. 

1969b. Studies on fossil phyllodont fishes : Casierius, a new genus of albulid from the 

Cretaceous of Europe and North America. Eclogae geol. Helv., Basle, 62 : 751-755, 2 pis. 

1969c. Two new late Cretaceous fishes from Montana and Wyoming. Breviora, Cam- 
bridge, Mass., 335 : 1-15, 4 pis. 

Fowler, H. W. 1936. The marine fishes of West Africa, Pt I. Bull. Am. Mus. nat. Hist., 

New York, 70 : vii 4- 605 pp., 275 figs. 
Fritsch, A. J. 1878. Die Reptilien und Fische der Bohmischen Kreideformation : ii + 44 pp., 

66 figs., 10 pis. Prague. 
Gardiner, B. G. 1963. Certain palaeoniscoid fishes and the evolution of the snout in actino- 

pterygians. Bull. Br. Mus. nat. Hist. (Geol.), London, 8 : 255-325, 2 pis., 20 figs. 

1970. Osteichthyes. In McGraw-Hill Yearbook of Science and Technology, ed. Lapedes : 

284-286, 1 fig. McGraw-Hill, New York. 

Garstang, W. 1931. The phyletic classification of the Teleostei. Proc. Leeds phil. lit. Soc. 

2, (5) : 240-260. 
Gaudant, J. 1968. Contribution a une revision des Anaethalion de Cerin (Ain). Bull. Bur. 

Rech. giol. min. Paris, (2) 4 : 95-115, 6 figs., 3 pis. 
Gill, T. 1893. Families and subfamilies of fishes. Mem. natn. Acad. Sci., Washington, 

4 : 125-138. 
Goody, P. C. 1969a. Sedenhorstia dayi (Hay), a new elopoid from the Cenomanian of Hajula 

in the Lebanon. Am. Mus. Novit., New York, 2358 : 1-23, 9 figs. 

1969b. The relationships of certain Upper Cretaceous teleosts with special reference to 

the myctophoids. Bull. Br. Mus. nat. Hist. (Geol.), London, Suppl., 7 : 1-255, I02 fig s - 

Gosline, W. A. i960. Contributions toward a classification of modern isospondylous fishes. 
Bull. Br. Mus. nat. Hist. (Zool.), London, 6 : 325-365, 15 figs. 

1961. Some osteological features of modern lower teleostean fishes. Smithson. misc. 

Collns., Washington, 142, 3 : 1-42, 8 figs. 

1965. Teleostean phylogeny. Copeia, Washington : 186-194, 1 fig. 

■ 1967. Reduction in branchiostegal ray number. Copeia, Washington : 237-239, 1 fig. 

1969. The morphology and systematic position of the alepocephaloid fishes. Bull. Br. 

Mus. nat. Hist. (Zool.), London, 18 : 183-218, 14 figs. 

1971. Functional morphology and classification of teleostean fishes. Univ. of Hawaii Press, 

Honolulu : ix + 208 pp., 29 figs. 
Greenwood, P. H. 1963. The swimbladder in African Notopteridae (Pisces) and its bearing 

on the taxonomy of the family. Bull. Br. Mus. nat. Hist. (Zool.), London, 5 : 377-412, 

5 figs., 4 pis. 

1966. The caudal fin skeleton in osteoglossoid fishes. Ann. Mag. nat. Hist., London, 

(13) 9: 581-597. 12 figs. 

■ 1968. The osteology and relationships of the Denticipitidae, a family of clupeomorph 

fishes. Bull. Br. Mus. nat. Hist. (Zool.), London, 16 : 213-273, 34 figs. 

1970a. Skull and swimbladder connections in fishes of the family Megalopidae. Bull. 

Br. Mus. nat. Hist. (Zool.), London, 19 : 119-135, 2 figs., 3 pis. 

1970b. On the genus Lycoptera and its relationship with the family Hiodontidae (Pisces, 

Osteoglossomorpha). Bull. Br. Mus. nat. Hist. (Zool.), London, 19 : 257-285, 10 figs. 



FOSSIL AND RECENT 217 

Greenwood, P. H., Myers, G. S., Rosen, D. E. & Weitzman, S. H. 1967. Named main 

divisions of teleostean fishes. Proc. biol. Soc. Wash., 80 : 227-228. 
Rosen, D. E., Weitzman, S. H. & Myers, G. S. 1966. Phyletic studies of teleostean 

fishes, with a provisional classification of living forms. Bull. Am. Mus. nat. Hist., New York, 

131 : 339-456, pis. 21-23, 32 charts. 
& Rosen, D. E. 1971. Notes on the structure and relationships of the alepocephaloid 

fishes. Am. Mus. Novit., New York, 2473: 1-41, 25 figs. 
Gregory, W. K. 1933. Fish skulls : a study of the evolution of natural mechanisms. Trans. 

Am. phil. Soc, Philadelphia, 23, 2 : 75-481, 299 figs. 
Gunther, A. C. 1877. Preliminary notes on new fishes collected in Japan during the expedi- 
tion of H.M.S. Challenger, Ann. Mag. nat. Hist., London, (4) 20 : 433-446. 
1887. Report on the deep-sea fishes collected by H.M.S. Challenger during the years 

1873-1876. Rep. scient. Results Voy. Challenger, London, 22 : 1-335, 73 pis- 
Harrisson, C. M. H. 1966. On the first halosaur leptocephalus, from Madeira. Bull. Br. 

Mus. nat. Hist. (Zool.), London, 14 : 441-486, 6 figs., 1 pi. 
Hay, O. P. 1903. On a collection of Upper Cretaceous fishes from Mount Lebanon, Syria, 

with description of four new genera and nineteen new species. Bull. Am. Mus. nat. Hist., 

New York, 19, 10 : 395-452, 14 pis. 
Hildebrand, S. F. 1963. Families Elopidae and Albulidae. In Fishes of the Western 

North Atlantic. Mem. Sears Fdn. mar. Res., New Haven, 1, 3 : xxi + 630 pp., 132 figs. 
Hilgendorf, F. 1877. Ptevothrissus, eine neve Clupeidengattung. Leopoldina, Dresden, 

13 : 127-128. 
Hollister, G. 1936. Caudal skeleton of Bermuda shallow water fishes. I. Order Isos- 

pondyli : Elopidae, Megalopidae, Albulidae, Clupeidae, Dussumieridae, Engraulidae. 

Zoologica, N.Y., 21 : 257-290, 53 figs. 

1939. Young Megalops cyprinoides from Batavia, Dutch East Indies, including a study of 

the caudal skeleton and a comparison with the Atlantic species, Tarpon atlanticus. Zoologica, 
N.Y., 24: 449-475, 21 figs. 

Holstvoogd, C. 1965. The pharyngeal bones and muscles in Teleostei, a taxonomic study. 

Proc. K. ned. Akad. Wet. Amsterdam, 68C : 209-218, 8 figs., 13 pis. 
Jessen, H. 1968. The gular plate and branchiostegal rays in Elops, A mia and Polypterus. In 

0rvig, T. (ed.), Nobel Symposium 4, Current problems of lower vertebrate phylogeny : 429- 

438, 3 figs. Stockholm. 
1972. Schultergurtel und Pectoralflosse bei Actinopterygien. Fossils and Strata, Oslo, 

1 : 1-101, pis. 1-25. 
Jordan, D. S. 1923. A classification of fishes, including families and genera as far as known. 

Stanford Univ. Pubis, Biol. Sci., Palo Alto, 3 : 79-243. 
& Evermann, B. W. 1896. The fishes of North and Middle America. Bull. U.S. natn. 

Mus., Washington, 16 : 1-1018. 
Lacepede, B. G. E. 1803. Histoire naturelle des Poissons, 5 (1798- 1803) : 803 pp. Paris. 
Laube, G. C. 1885. Eine Beitrage zurKenntnissder Fische des bohmischenTurons. Denkschr. 

Akad. Wiss. Wien. 1 : 258-298. 
Linnaeus, K. 1758. Sy sterna Naturae, 10th ed. : 824 pp. Stockholm. 

1766. Systema Naturae, 12th ed. : 1327 pp. Stockholm. 

Loomis, F. 1900. Die Anatomie und die Verwandtschaft der Ganoid und Knochen Fische 

aus der Kreide Formation von Kansas. Palaeontographica, Stuttgart, 46 : 213-284, 

pis. 19-27. 
McAllister, D. E. 1968. Evolution of branchiostegals and classification of teleostome fishes. 

Bull. natn. Mus. Can., Ottawa, 221 : xiv 4- 239 pp., 21 pis. 
McDowell, S. In press. Order Heteromi. In Fishes of the Western North Atlantic. Mem. 

Sears Fdn. mar. Res., New Haven. 
Mantell, G. A. 1822. The fossils of the South Downs ; or Illustrations of the Geology of Sussex: 

320 pp., 42 pis. Lupton Relfe, London. 
Mantell, G. A. 1838. The Wonders of Geology : xvi + 828 pp. Relfe & Fletcher, London. 



218 ELOPIFORM FISHES 

von der Marck, \V. 1858. Ueber einige Wirbelthiere, Kruster und Cephalopoden der West- 
phalischen Kreide. Z. dt. geol. Ges., Berlin, 10 : 231-271. 

1863. Fossile Fische, Krebse und Pflanzen aus dem Plattenkalk der Jiingsten Kreide in 

Westphalen. Palaeontographica, Cassel, 11 : 1-83, 14 pis. 

1868. Neue Fische und Krebse und Pflanzen aus dem Plattenkalk der Jiingsten Kreide 

in Westphalen. Palaeontographica, Cassel, 15 : 269-305, 4 pis. 

1873. Neue beitrage zur Kenntniss der fossilen Fische und anderer Thierreste aus der 

Jiingsten Kreide Westfalens. Palaeontographica, Cassel, 22 : 55-74, 2 pis. 

1885. Fische der Oberen Kreide Westfalens. Palaeontographica, Cassel, 31 : 233-267, 

5 Pis- 
Monod, T. 1967. Le complexe urophore des Teleosteens : typologie et evolution (note 

preliminaire) . Collogues int. Cent. natn. Res. scient., Paris, 163 : 111-131, 16 figs. 
1968. Le complexe urophore des poissons teleosteens. MSm. Inst. Fond. Afr. noire, 

Dakar, 81 : 1-705, 989 figs. 
Munster, G. 1842a. Beitrag zur Kenntniss einiger neuen seltenen Versteinerungen aus den 

lithographischen Sheifern in Baiern. Neues Jb. Miner. Geogn. Geol. Petrefakt., Stuttgart : 

35-46. 
1842b. Beschreibung einiger neuen Fische aus den lithographischen Scheifern von 

Bayern. Beitrage zur Petrefactenkunde, Bayreuth, 5 : 55-64, 4 pis. 
Nelson, G. J. 1967. Gill arches of teleostean fishes of the family Clupeidae. Copeia, 

Washington : 389-399, 9 figs. 

1968a. Gill arches of teleostean fishes of the division Osteoglossomorpha. /. Linn. Soc. 

(Zool.), London, 47 : 261-277, 11 figs. 

1968b. Gill arch structure in Acanthodes. In 0rvig, T. (ed.), Nobel Symposium 4, 

Current problems of lower vertebrate phytogeny : 129-143, 6 figs. Stockholm. 

1969a. Gill arches and the phylogeny of fishes, with notes on the classification of ver- 
tebrates. Bull. Am. Mus. nat. Hist., New York, 141 : 475-552, 26 figs., 13 pis. 

1969b. Infraorbital bones and their bearing on the phylogeny and geography of osteo- 

glossomorph fishes. Am. Mus. Novit., New York, 2394 : 1-37, 22 figs. 
Norman, J. R. i960. A draft Synopsis of the Orders, Families and Genera of Recent Fishes and 

Fish-like Vertebrates : 1-649. Br. Mus. (Nat. Hist.), London. 
Nursall, J. R. 1964. The jaws of pycnodonts (Holostei : Pycnodontiformes), Am. Zool., 

New York, 4 : 75. 
Nybelin, O. 1956. Les canaux sensoriels du museau chez Elops saurus. Ark. Zool., Stockholm, 

(2) 10, 9:453-458, 3 figs. 

i960. A gular plate in Albula vulpes (L.). Nature, Lond. 188 : 78. 

1963. Zur Morphologie und Terminologie des Schwanzskelettes der Actinopterygier. 

Ark. Zool., Stockholm, (2) 15 : 485-516, 22 figs. 

1964. Versuch einer taxonomischen Revision der jurassischen Fisch-gattung Thrissops 

Agassiz. Goteborgs K. Vetensk.- o. Vitterh.-Samh. Handl., (B) 9 : 1-44, 9 pis. 

1967a. Notes on the reduction of the sensory canal system and of the canal-bearing 

bones in the snout of higher actinopterygian fishes. Ark. Zool., Stockholm, (2) 19 : 235-246, 
4 figs. 

1967b. Versuch einer taxonomischen Revision der Anaethalion- Arten des Weissjura 

Deutschlands. Acta R. Soc. scient. litt. gothoburg., 2 : 1-53, 5 figs., 1 table, 8 pis. 

1968. The dentition in the mouth cavity of Elops. In 0rvig, T. (ed.), Nobel Symposium 

4, Current problems of lower vertebrate phylogeny : 439-443, 3 figs. Stockholm. 

1971. On the caudal skeleton in Elops with remarks on other teleostean fiishes. Acta 

R. Soc. scient. litt. gothoburg, 7 : 1-52, 12 pis. 

Okada, Y. i960. Freshwater fishes of Japan. /. Fac. Fish. pref. Univ. Mie-Tsu, 4 (2) : 267- 
588, 68 figs., 53 pis., 89 tables. 

Owen, R. 1840. Odontography, London, 1840-1845 : lxxiv + 655 pp., 168 pis. 

1854. Descriptive Catalogue of the Fossil Organic Remains of Reptilia and Pisces con- 
tained in the Museum of the Royal College of Surgeons of England : xix + 184 pp. London. 



FOSSIL AND RECENT 219 

Patterson, C. 1967a. Are the teleosts a polyphyletic group ? Colloques int. Cent. natn. 
Res. sclent., Paris, 163 : 93-109, n figs. 

1967b. Teleostei. In Harland, W. B. et al., The Fossil Record : 654-666. Geological 

Society, London. 

1967c. A second specimen of the Cretaceous teleost Protobrama and the relationships of 

the sub-order Tselfatioidei. Ark. Zool., Stockholm, (2) 19, 8 : 215-234, 8 figs. 

1968a. The caudal skeleton in Lower Liassic pholidophorid fishes. Bull. Br. Mus. nat. 

Hist. (Geol.), London, 16 : 201-239, 5 pis., 12 figs. 

1968b. The caudal skeleton in Mesozoic acanthopterygian fishes. Bull. Br. Mus. nat. 

Hist. (Geol.), London, 17, 2 : 47-102, 28 figs. 

1970a. A clupeomorph fish from the Gault (Lower Cretaceous). /. Linn. Soc. (Zool.), 

London, 49 : 161 -182, 5 figs., 1 pi. 

1970b. Two Upper Cretaceous salmoniform fishes from the Lebanon. Bull. Br. Mus. 

nat. Hist. (Geol.), London, 19 : 205-296, 48 figs., 5 pis. 

Pictet, F. J. 1850. Description de quelques poissons fossiles du Mont Liban, 59 pp., 10 pis. 

J.-G. Fick, Geneve. 
& Humbert, A. 1866. Nouvelles recherches sur les poissons fossiles du Mont Liban, 

vii + 114 pp., 19 pis. Georg, Geneve. 
Poll, M. 1953. Poissons III. Teleosteens malacopterygiens. Re's. set. exped. ocianogr. 

Belg. eaux cot. Afr. atlant. sud. (ig48-iQ4g), 4, 2 : 1-258, 8 pis., 104 figs. 
Priem, F. 1908. Etudes des poissons fossiles du Bassin Parisien. Annls Paleont., Paris, 

3 : 1 -144, 5 Pis-. 74 figs. 
Regan, C. T. 1909. A revision of the fishes of the genus Elops. Ann. Mag. nat. Hist., London, 

(8) 3 : 37-40. 

1910. The caudal fin of the Elopidae and some other teleostean fishes. Ann. Mag. nat. 

Hist., London, (8) 5 : 354-358, 2 figs. 

Ridewood, W. G. 1904. On the cranial osteology of the fishes of the families Elopidae and 

Albulidae, with remarks on the morphology of the skull in the lower teleostean fishes 

generally. Proc. Zool. Soc. Lond., 2 : 35-81, 18 figs. 
Romer, A. S. 1966. Vertebrate Palaeontology, 3rd ed. : viii : 468 pp., 443 figs., 4 tables. 

Chicago University Press. 
Rosen, D. E. & Greenwood, P. H. 1970. Origin of the Weberian apparatus and relationship 

of the ostariophysan and gonorynchiform fishes. Am. Mus. Novit., New York, 2428 : 1- 

25, 16 figs. 
& Patterson, C. 1969. Thestructure and relationships of the paracanthopterygian fishes. 

Bull. Am. Mus. nat. Hist., New York, 141 : 357-474, 74 figs., pis. 52-78, 8 tables. 
Saint-Seine, P. de. 1949. Les poissons des Calcaires lithographiques de Cerin (Ain). Nouv. 

Archs Mus. Hist. nat. Lyon, 2 : vii + 357 pp., 26 pis. 
Sauvage, H. E. 1891. Note sur quelque poissons fossiles du Lias superieur de l'Yonne. Bull. 

Soc. Sci. hist. nat. Yonne, Auxerre, 45 (2) : 31-45, 3 pis. 

1903. Noticia sobre los Peces de la Caliza litografica de la Provincia de Lerida (Cataluna). 

Mems R. Acad. Cienc. Artes Barcelona, 3 (4), no. 35 : 1-32, 4 pis. 

Scopoli, G. A. 1777. Introductio ad historiam naturalem sistema genera lapidum, plantarum 

et animulium : 6 + 506 + 34 pp. Prague. 
Siegfried, P. 1954. Di e Fisch-Fauna des Westfalischen Ober-Senons. Palaeontographica, 

Stuttgart, (A) 106 : 1-36, 15 pis., 2 figs. 
Starks, E. C. 1930. The primary shoulder girdle of the bony fishes. Stanford Univ. Pubis, 

Biol. Sci., Palo Alto, 6 : 3-93, 38 figs. 
Stewart, A. 1898. A contribution to the knowledge of the ichthyic fauna of the Kansas 

Cretaceous. Kans. Univ. Q., Lawrence, 7 : 21-29, 2 pis. 
Taylor, W. R. 1967. An enzyme method of clearing and staining small vertebrates. Proc. 

U.S. natn. Mus., Washington, 122 : 1-17. 
Tchernavin, V. 1953. The Feeding Mechanisms of a Deep Sea Fish , Chauliodus sloani Schneider : 

99 pp., 10 pis., 36 figs. Br. Mus. (Nat. Hist.), London. 



220 ELOPIFORM FISHES 

Vetter, B. 1878. Untersuchungen zur vergleichenden Anatomie der Kiemen- und Kiefer- 

musculatur der Fische. Jena Z. Naturw., 12 : 431-450, 3 pis. 
Vrba, E. S. 1968. Contributions to the functional morphology of fishes. Part V. The 

feeding mechanisms of Elops saurus Linnaeus. Zoologica Afr., Cape Town, 3 : 211-236, 

9 figs. 
Wagner, A. 1863. Monographic der fossilen Fische aus den lithographischen Schiefern 

Bayerns. Zweite Abtheilung. Abh. bayer. Akad. Wiss., Miinchen, 9 : 611-748, 6 pis. 
Wenz, S. 1965. Lespoissons Albiensde Vallentigny (Aube). Annls PaUont., Paris, 51 : 1-24, 

5 figs., 2 pis. 
White, E. I. 1938. The generic name Aethalion. Ann. Mag. nat. Hist., London, (10) 19 : 319. 

i960. British Caenozoic Fossils, 2nd ed. : i-vii + 132 pp., 44 pis. Brit. Mus. (Nat. 

Hist.), London. 

& Frost, G. A. 193 1. The Vertebrate Faunas of the English Eocene. Vol. 1. From the 

Thanet Sands to the Basement Bed of the London Clay: 1-123, 1 pi. Brit. Mus. (Nat. 

Hist.), London. 
& Moy-Thomas, J. A. 1940. Notes on the nomenclature of fossil fishes — Part III, 

Homonyms M-Z. Ann. Mag. nat. Hist., London, (2) 7 : 395-400. 
Whitehead, P. J. P. 1962. The species of Elops (Pisces : Elopidae). Ann. Mag. nat. Hist., 

London, (13) 5 : 321-329, 3 figs. 

1963. A contribution towards the classification of clupeoid fishes. Ann. Mag. nat. Hist., 

London, (13) 5 : 737-75°. 3 figs- 

Winkler, T. C. 1862. Description de quelques nouvelles especes des poissons fossiles du 

calcaire lithographique de Solenhofen. Natuurk. Verh. holland Maatsch. Wet. Haarlem, 

16 : 94 pp., 10 pis. 
Woodward, A. A. & Sherborn, C. D. 1890. A Catalogue of British Fossil Vertebrata : xxxv + 

396 pp. Dulau & Co., London. 
Woodward, A. S. 1888. A synopsis of the vertebrate fossils of the English Chalk. Proc. 

Geol. Ass., London, 10 : 273-338, 1 pi. 
1 89 1. Notes of some fish remains from the Lower Tertiary and Upper Cretaceous of 

Belgium, collected by Monsieur A. Houzeau de Lehaie. Geol. Mag., London, 8 : 104-114, 

1 pi. 
1893. Description of the skull of Pisodus owenii, an Albula-like fish of the Eocene period. 

Ann. Mag. nat. Hist., London (6) 11 : 357-359, 1 pi. 

1895. Catalogue of the Fossil Fishes in the British Museum {Natural History) : 3 : xlii + 544 

pp., 18 pis., 45 figs. Brit. Mus. (Nat. Hist.), London. 

1900. On a new specimen of the clupeoid fish Aulolepis typus from the English Chalk. 

Ann. Mag. nat. Hist., London, (7) 5 : 324-326, 1 pi. 

1901. Catalogue of the Fossil Fishes in the British Museum [Natural History), 4, xxxviii + 

636 pp., 19 pis., 22 figs. Brit. Mus. (Nat. Hist.), London. 

1907. The fossil fishes of the English Chalk, pt. 3. Palaeontogr. Soc. (Monogr.), London : 

97-128, 6 pis., 14 figs. 

1908. The fossil fishes of the English Chalk, pt. 4. Palaeontogr. Soc. (Monogr.), London : 

129-152, 5 pis., 6 figs. 

1919. The fossil fishes of the English Wealden and Purbeck Formations, Pt. 3. Palaeontogr. 

Soc. (Monogr.), London : 105-148, 6 pis., 6 figs. 

1942. Some new and little known Upper Cretaceous fishes from Mount Lebanon. Ann. 

Mag. nat. Hist., London, (11) 9 : 537-568, 4 figs., 5 pis. 

Yabe, H. & Okada, T. 1930. On some fossil fishes from the Cretaceous of Japan. Jap. J. 
Geol. Geogr., Tokyo, 8 : 1-7, 2 pis. 



FOSSIL AND RECENT 
VIII. ABBREVIATIONS USED IN TEXT-FIGURES 



ace 


anterior ceratohyal 


fm 


foramen magnum 


ao 


antorbital 


foa 


foramen for orbital artery 


ao.s.c 


antorbital sensory canal 


f.ptf 


foramen leading to 


apal 


autopalatine 




post-temporal fossa 


a.p.b 


anterior opening for periotic 


fr 


frontal 




bulla 


fsp 


foramen for occipital nerve 


a.ros 


anterior rostral ossicle 


fuv 


fused vertebral centrum 


art 


articular 


gr.ah 


groove for afferent hyoidean 


asp 


autosphenotic 




artery 


atfc 


anterior opening of pars 


gr.V & VII 


groove for superficial 




jugularis 




ophthalmic branches of 


bh 


basihyal 




V& VII 


boc 


basioccipital 


gr.so.s.c 


groove for supraorbital sensory 


brr 


branchiostegal ray 




canal 


bsp 


basisphenoid 


gu 


gular plate 


cb 


ceratobranchial (numbered 1-5) 


h 


hypural (numbered 1-7) 


cl 


cleithrum 


hb 


hypobranchial (numbered 1-3) 


cor 


coracoid 


hm 


hyomandibular 


CSC 


caudal scute 


ib 


infrapharyngobranchial 


cy 


scale 




(numbered 1-3) 


de 


dermethmoid 


ic 


intercalar 


den 


dentary 


in 


interhyal 


df 


dilatator fossa 


io 


infraorbital (numbered 1-5) 


dhh 


dorsal hypohyal 


iop 


interoperculum 


d.l 


dorsal (epiotic) limb of 


io.s.c 


infraorbital sensory canal 




post-temporal 


l.e 


lateral ethmoid 


dpal 


dermopalatine 


1.1 


lateral line 


dsp 


dermosphenotic 


m.c 


Meckelian cartilage 


eart 


endosteal articular 


m.cor 


mesocoracoid 


eb i 


first epibranchial 


mes 


mesethmoid 


e.com 


ethmoid commissure 


mpt 


metapterygoid 


ecp 


ectopterygoid 


m.ros 


middle rostral ossicle 


enp 


endopterygoid 


m.s.c 


mandibular sensory canal 


ep 


epural (numbered 1-3) 


mx 


maxilla 


epo 


epiotic 


myp 


posterior myodome 


epo.pr 


epiotic process 


na 


nasal 


exo 


exoccipital 


np 


notochordal pit 


f.ah 


foramen for afferent hyoidean 


npu 1 


neural arch associated with 




artery 




first preural centrum 


fahm 


facet for articulation with 


nsp pu 


neural spine associated with 




hyomandibular 




preural centrum 


fa.mx 


facet for articulation with 




(numbered 2-4) 




maxilla 


nu 1 


neural arch associated with 


fa. pal 


facet for articulation with 




first ural centrum 




palatine 


op 


operculum 


fa.pmx 


facet for articulation with 


ors 


orbitosphenoid 




premaxilla 


ot.s.c 


otic sensory canal 


f.bv 


foramen for a blood vessel 


pa 


parietal 


fep 


foramen for efferent 


pap 


parapophysis 




pseudobranchial artery 


par 


parasphenoid 


fhv 


foramen for head vein 


pee 


posterior ceratohyal 


fica 


foramen for internal carotid 


pel 


postcleithrum (numbered 1-3) 




artery 


ph 


parhypural 



ELOPIFORM FISHES 



pmx 


premaxilla 


soc 


pop 


preoperculum 


sop 


pop.s.c 


preopercular sensory canal 


so.s.c 


p.p.b 


posterior opening of periotic 


stt 




bulla 


stt.com 


pr.l.e 


lateral ethmoid process of 


suf 




autopalatine 


sy 


pr.mes 


mesethmoid process of 
autopalatine 


Tpbb 


pro 


prootic 


Tpbh 


pro.ic 


prootic-intercalar bridge 


Tpce 


p.ros 


posterior rostral ossicle 


Tpib4 


psp 


pterosphenoid 




ptf 


post-temporal fossa 


u 


ptfc 


posterior opening of pars 


ud 




jugularis 


un 


pto 


pterotic 


vhh 


ptt 


post-temporal 


v.l 


pu 


preural centrum 






(numbered 1-6) 


vo 


qu 


quadrate 


I 


r.d 


distal radial 


II 


ros 


rostral ossicles 


III 


r.p 


proximal radial 
(numbered 1-4) 


Vpal 


sb i 


first suprapharyngobranchial 


VII hm 


sea 


scapula 




scaf 


scapular foramen 


VII m 


scl 


supraclei thrum 




sef 


sub-epiotic fossa 


VII ot 


ses 


sesamoid articular 




smx 


supramaxilla (numbered 1-2) 


IX 


so 


supraorbital 


X 



supraoccipital 

suboperculum 

supraorbital sensory canal 

supratemporal 

supratemporal commissure 

subtemporal fossa 

symplectic 

tooth plate on basibranchial 

(numbered 1-4) 
tooth plate on basihyal 
tooth plate on ceratohyal 
tooth plate on 

infrapharyngobranchial 4 
ural centrum (numbered 1-2) 
urodermal 

uroneural (numbered 1-7) 
ventral hypohyal 
ventral (intercalar) limb of 

post-temporal 
vomer 

foramen for olfactory tract 
foramen for optic tract 
foramen for oculomotor 
foramen for palatine ramus of 

trigeminal 
foramen for hyomandibular 

trunk of facial 
foramen for mandibular ramus 

of facial 
foramen for otic branch of 

facial 
foramen for glossopharyngeal 
foramen for vagus 



Peter L. Forey 
Department of Biology 
Queen Elizabeth College 
University of London 
London 



A LIST OF SUPPLEMENTS 

TO THE GEOLOGICAL SERIES 

OF THE BULLETIN OF 

THE BRITISH MUSEUM (NATURAL HISTORY) 






i. Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. 
Pp. 213 ; 30 Plates ; 2 Text-figures. 1965. £6. 

2. El-Naggar, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper 
Cretaceous — Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, 
Egypt, U.A.R. Pp. 291 ; 23 Plates ; 18 Text-figures. 1966. £10. 

3. Davey, R. J., Downie, C, Sarjeant, W. A. S. & Williams, G. L. Studies on 
Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248 ; 28 Plates ; 64 Text- 
figures. 1966. £y. 

3. Appendix. Davey, R. J., Downie, C, Sarjeant, W. A. S. & Williams, G. L. 
Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 
1969. 8op. 

4. Elliott, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of 
the Middle East. Pp. in ; 24 Plates ; 17 Text-figures. 1968. &5.I2J. 

5. Rhodes, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni- 
ferous) Conodont faunas, and their value in local and continental correlation. 
Pp. 315 ; 31 Plates ; 92 Text-figures. 1969. £11. 

6. Childs, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern 
Europe. Pp. 119 ; 12 Plates ; 40 Text-figures. 1969. £4.75. 

7. Goody, P. C. The relationships of certain Upper Cretaceous Teleosts with 
special reference to the Myctophoids. Pp. 255 ; 102 Text-figures. 1969. 
£6.50. 

8. Owen, H. G. Middle Albian Stratigraphy in the Paris Basin. Pp. 164 ; 
3 Plates ; 52 Text-figures. 1971. £6. 

9. Siddiqui, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae 
from West Pakistan. Pp. 98 ; 42 Plates ; 7 Text-figures. 1971. £8. 



Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Prets, Bristol BS4 }NU 












ORDOVICIAN BRACHIOPODA 

FROM THE 

SHELVE DISTRICT, SHROPSHIRE 



A. WILLIAMS 



BULLETIN OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 
GEOLOGY Supplement n 

LONDON: 1974 



': 



"6 DEC 191 



ORDOVICIAN BRACHIOPODA FROM THE ^SjijggS 
SHELVE DISTRICT, SHROPSHIRE 



BY 

ALWYN WILLIAMS 

The University of Birmingham 



Pp. i -163 ; 28 Plates ; 11 Text-figures ; no Tables 



BULLETIN OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 
GEOLOGY Supplement 11 

LONDON: 1974 



THE BULLETIN OF THE BRITISH MUSEUM 

(natural history), instituted in 1949, is 
issued in five series corresponding to the Departments 
of the Museum, and an Historical series. 

Parts will appear at irregular intervals as they 
become ready. Volumes will contain about three or 
four hundred pages, and will not necessarily be 
completed within one calendar year. 

In 1965 a separate supplementary series of longer 
papers was instituted, numbered serially for each 
Department. 

This paper is Supplement 11 of the Geological 
(Palaeontological) series. The abbreviated titles of 
periodicals cited follow those of the World List of 
Scientific Periodicals. 



World List abbreviation : 
Bull. Br. Mus. nat. Hist. (Geol.) Suppl. 



Trustees of the British Museum (Natural History), 1974 



TRUSTEES OF 
THE BRITISH MUSEUM (NATURAL HISTORY) 

Issued 2 December 1974 Price £12.80 



ORDOVICIAN BRACHIOPODA FROM THE 
SHELVE DISTRICT, SHROPSHIRE 

By ALWYN WILLIAMS 



CONTENTS 










Page 


I. Introduction ....... . 8 


II. Faunal distribution 










9 


III. Faunal associations 










16 


IV. Systematic methods 










23 


V. Systematic palaeontology . 










25 


Class INARTICULATA Huxley 










25 


Order Lingulida Waagen . 










25 


Superfamily Lingulacea Menke 










25 


Family Obolidae King 










25 


Subfamily Obolinae King 










25 


Obolus Eichwald . 










25 


Obolus subditivus sp. nov. 










25 


Obolus sp. 










26 


Schmidtites Schuchert & Le Vene 








26 


Schmidtites ? simplex sp. nov. 








26 


Schmidtites ? simplex subcircularis sp. e 


t subsp. no> 


T. 


27 


Subfamily Lingulellinae Schuchert 






28 


Lingulella Salter .... 






28 


Lingulella displosa sp. nov. . 






28 


Lingulella displosa petila sp. et subsp. i 


lOV. . 




29 


Palaeoglossa Cockerell . 






31 


Palaeoglossa attenuata Sowerby . 






32 


Palaeoglossa myttonensis sp. nov. . 






33 


Subfamily Glossellinae Cooper 






35 


Pseudolingula Mickwitz 






35 


Pseudolingula spatula sp. nov. 






36 


Family Elkaniidae Walcott & Schuchert 






38 


Monobolina Salter 






38 


Monobolina plumbea (Salter) 






38 


Family Paterulidae Cooper 






40 


Paterula Barrande 






40 


Paterula cf. bohemica Barrande . 






40 


Paterula cf. perfecta Cooper 






4i 


Order Acrotretida Kuhn . 






42 


Suborder Acrotretidina Kuhn . 






42 


Superfamily Acrotretacea Schuchert . 






42 


Family Acrotretidae Schuchert 






42 


Apsotreta Palmer 






42 


Apsotreta sp. .... 






42 


Conotreta Walcott 






43 


Conotreta stapeleyensis sp. nov. 






43 


Superfamily Discinacea Gray 










44 



SHELVE DISTRICT 



Family Trematidae Schuchert . 
Schizocrania Hall & Whitfield 
Schizocrania salopiensis sp. nov, 
Family Discinidae Gray . 
Subfamily Orbiculoideinae Schuchert & Le Vene 
Orbiculoidea d'Orbigny 

Orbiculoidea sp. . 
Schizotreta Kutorga 
Schizotreta transversa sp. nov 
Schizotreta sp. . 
Suborder Craniidina Waagen 
Superfamily Craniacea Menke 
Family Craniidae Menke . 
Petrocrania Raymond . 
Petrocrania dubia sp. nov. 
Class ARTICULATA Huxley 
Order Orthida Schuchert & Cooper 
Suborder Orthidina Schuchert & Cooper 
Superfamily Orthacea Woodward 
Family Hesperonomiidae Ulrich & Cooper 
Hesperonomia Ulrich & Cooper 
Hesperonomia sp. 
Family Orthidae Woodward 
Subfamily Orthinae Woodward 
Orthis Dalman .... 

Orthis cf. callactis Dalman . 
Orthis sp. ..... 

Lenorthis Andreeva 

Lenorthis cf. proava (Salter) 
Orthambonites Pander . 
Orthambonites exopunctata sp. nov. 
Subfamily Productorthinae Schuchert & Cooper 
Nicolella Reed 
Nicolella cf. actoniae (J. de C. Sowerby) 
Subfamily Whittardiinae nov. 
Whittardia gen. nov. 
Whittardia paradoxica gen. et sp. nov. 
Family Dolerorthidae Opik 
Subfamily Dolerorthinae Opik 
Dolerorthis Schuchert & Cooper 
Dolerorthis cf . tenuicostata Williams 
Subfamily Glyptorthinae Schuchert & Cooper 
Glyptorthis Foerste 
Glyptorthis viriosa sp. nov. . 
Family Alimbellidae Andreeva 
Astraborthis gen. nov. . 
Astraborthis uniplicata gen. et sp. nov. 
Family Finkelnburgiidae Schuchert & Cooper 
Diparelasma Ulrich & Cooper 
Diparelasma sp. 
Family Plectorthidae Schuchert & Le Vene 
Subfamily Plectorthinae Schuchert & Le Vene. 
Plectorthis Hall & Clarke 
Plectorthis whitteryensis sp. nov. 



44 
44 
44 

47 
47 
47 
47 
47 
47 
48 
48 



48 
48 
49 
49 
49 
49 
49 
49 
49 
50 
50 
50 
50 
51 
52 
52 
53 
53 
57 
57 
57 
60 
62 
62 
63 
63 
63 
63 
64 
64 
64 
68 
68 
70 
7i 
7i 
7i 
7i 
7i 
7i 
7i 



ORDOVICIAN BRACHIOPODA 



Plectorthis sp. . 
Desmorthis Ulrich & Cooper . 

Desmorthis ? sp. nov. . 
Gelidorthis Havlicek 

Gelidorthis cf. partita (Barrande) . 
Tazzarinia Havlicek 

Tazzarinia elongata sp. nov. 
Subfamily Platystrophiinae Schuchert & Le Vene 
Platystrophia King .... 

Platystrophia caelata sp. nov. 

Platystrophia cf . major Williams . 
Mcewanella Foerste .... 

Mcewanella sp. . 
Salacorthis gen. nov. .... 

Salacorthis costellata gen. et sp. nov. 
Family Skenidiidae Kozlowski. 
Skenidioides Schuchert & Cooper . 

Skenidioides cf. costatus Cooper . 
Protoskenidioides gen. nov. . 

Pvotoskenidioides revelata gen. et sp. nov. 
Superfamily Enteletacea Waagen 
Family Schizophoriidae Schuchert & Le Vene 
Subfamily Draboviinae Havlicek 
Drabovia Havlicek .... 

Drabovia cf . fascicostata Havlidek 
Nocturniella Havlicek .... 

Nocturniella sp. 
Family Dalmanellidae Schuchert 
Dalmanella Hall & Clarke 

Dalmanella parva Williams . 

Dalmanella salopiensis sp. nov. . 

Dalmanella salopiensis gregaria sp. et subsp. nov 

Dalmanella salopiensis transversa sp. et subsp. nov 

Dalmanella elementaria sp. nov. . 
Onniella Bancroft 

Onniella ostentata Williams lepida subsp. nov 
Family Harknessellidae Bancroft 
Harknessella Reed 

Harknessella cf. subplicata Bancroft 
Horderleyella Bancroft. 

Horderleyella cf. plicata Bancroft. 

Horderleyella sp. 
Reuschella Bancroft 

Reuschella horderleyensis Bancroft carinata subsp 
Family Heterorthidae Schuchert & Cooper 
Heterorthis Hall & Clarke 

Heterorthis sp. .... 
Tissintia Havlicek 

Tissintia prototypa (Williams) 

Tissintia immatura (Williams) 
Family Linoporellidae Schuchert & Cooper 
Salopia Williams 

Salopia cf. salteri (Davidson) 

Salopia sp. .... 



73 
73 
73 
74 
74 
75 
75 
76 
76 
76 
78 
78 
78 
79 
80 
82 
82 
82 
83 
85 
87 
87 
87 
87 
87 
88 
88 
89 
89 
89 
90 

9i 
92 

97 
98 
98 
102 
102 
102 
102 
102 
104 
104 
104 
106 
106 
106 
107 
108 
109 
114 
114 
114 
114 



SHELVE DISTRICT 



Suborder Clitambonitidina Opik 
Superfamily Gonambonitacea Schuchert & Cooper 
Family Gonambonitidae Schuchert & Cooper 
Subfamily Anomalorthinae Ulrich & Cooper 
Oslogonites Opik 
Oslogonites ? sp. 
Family Kullervoidae Opik 
Kullervo Opik 
Kullervo sp. 
Suborder Triplesiidina Moore . 
Superfamily Triplesiacea Schuchert 
Family Triplesiidae Schuchert . 
Triplesia Hall 

Triplesia sp. . 
Bicuspina Havlicek 
Bicuspina subquadrata sp. nov. 
Bicuspina modesta sp. nov. . 
Caeroplecia gen. nov. . 

Caeroplecia plicata gen. et sp. nov 
Oxoplecia Wilson 
Oxoplecia cf. nantensis MacGregor 
Order Strophomenida Opik 
Suborder Strophomenidina Opik 
Superfamily Plectambonitacea Jones 
Family Leptestiidae Opik 
Subfamily Leptestiinae Opik 
Palaeostrophomena Holtedahl 
Palaeostrophomena sp. 
Family Eocramatiidae nov. 
Eocramatia gen. nov. . 
Eocramatia dissimulata gen. et sp. nov. 
Family Sowerbyellidae Opik 
Subfamily Sowerbyellinae Opik 
Sowerbyella Jones 
Sowerbyella antiqua Jones 
Sowerbyella cf. antiqua Jones 
Sowerbyella multiseptata sp. nov. 
Sowerbyella cf . sericea permixta Williams 
Sowerbyella cf. sericea (J. de C. Sovverby) 
Subfamily Aegiromeninae Havlicek 
Sericoidea Lindstrom . 
Sericoidea cf. abdita Williams 
Superfamily Strophomenacea King 
Family Strophomenidae King . 
Subfamily Furcitellinae Williams . 
Furcitella Cooper 
Furcitella sp. .... 
Murinella Cooper 
Murinella sp. . 
Subfamily Rafinesquininae Schuchert 
Rafinesquina Hall & Clarke . 
Rafinesquina delicata sp. nov. 
Rafinesquina sp. 
Kjaerina (Hedstroemina) Bancroft 



:i5 

15 
15 
15 
:i5 
ti5 
[15 
115 
[16 
:i6 
[16 
ri6 
:i6 
[17 

:i7 

:i8 

[21 
[22 
[2 4 

125 
[26 
[26 
[26 
[26 
[26 
[26 
[26 

[27 
[28 

[29 

[30 
[30 
130 
[30 
131 
[32 

[33 
134 
[39 
C39 
[39 
[40 
[40 
[40 
[40 
[40 
[41 
[41 
[41 
[41 
[41 
[43 
[45 



ORDOVICIAN BRACHIOPODA 



Kjaerina (Hedstroemina) sp. 
Subfamily Glyptomeninae Williams 
Bystromena gen. nov. . 
Bystromena perplexa gen. et sp. nov. 
Family Leptaenidae Hall & Clarke . 
Leptaena Dalman 

Leptaena cf . ventricosa Williams . 
Kiaeromena Spjeldnaes 
Kiaeromena cf. kjerulfi (Holtedahl) 
Order Pentamerida Schuchert & Cooper 
Suborder Syntrophiidina Ulrich & Cooper 
Superfamily Porambonitacea Davidson 
Family Syntrophiidae Schuchert 
Subfamily Xenelasmatinae Ulrich & Cooper 
Euorthisina Havlicek . 
Euorthisina cf . moesta minor Havlicek 
Family Parastrophinidae Ulrich & Cooper 
Parastrophinella Schuchert & Cooper 
Parastrophinella musculosa sp. nov. 
Parastrophinella sp. . 
Order Rhynchonellida Kuhn 

Superfamily Rhynchonellacea Gray 
Family Trigonirhynchiidae McLaren 
Rostricellula Ulrich & Cooper 
Rostricellula sparsa Williams 
Order Spiriferida Waagen . 
Suborder Atrypidina Moore 
Superfamily Atrypacea Gill 
Family Atrypidae Gill 
Subfamily Zygospirinae Waagen 
Zygospira Hall 
Zygospira sp. 
VI. Acknowledgments 
VII. References . 
VIII. Index 



145 
J 45 
145 
146 
148 
148 
148 
149 
149 
150 
150 
150 
150 
150 
150 
150 
151 
151 
151 
152 
153 
153 
153 
153 
153 
154 
154 
154 
154 
!54 
154 
!54 
154 
155 
158 



SYNOPSIS 

A study of fossil Brachiopoda collected mainly as moulds from the Ordovician successions of 
the Shelve area, reveals the presence of 83 species and subspecies (35 of them new) belonging to 
63 genera, of which Astraborthis, Protoskenidioides , Salacorthis and Whittardia are new orthaceans, 
Caeroplecia a new triplesiacean, Eocramatia a new plectambonitacean and Bystromena a new 
strophomenacean. Whittardia is placed in a new subfamily Whittardiinae and Eocramatia in 
a new family Eocramatiidae. 

The number of taxa identified in assemblages recovered from any one formation is small, and 
only exceeds 20 in the Spy Wood Grit and Whittery Shales. This relative poverty of the faunas 
is at least partly related to rock type, with a dominantly inarticulate association comprising 
only a few genera prevalent in shales, and a richer, mainly articulate association more typical 
of the siltstones and sandstones. Both associations evolved by significant replacements of 
genera as well as species from one horizon to another, so that no combination of taxa can be 
described as consistently characteristic of either. 

The faunas are also restricted in geographical distribution. They show affinities with assem- 
blages from Wales and E. Shropshire and confirm the currently adopted correlation of the 



8 SHELVE DISTRICT 

Shelve rocks with the standard Anglo-Welsh succession. However, taxa which also occur in 
Bohemia and Morocco generally appear earlier in the Shropshire successions and do not assist 
in precise correlation with either region. 



I. INTRODUCTION 

This paper is primarily a systematic study of Ordovician brachiopods collected by 
the late Professor W. F. Whittard with the assistance of Mr T. R. Fry during an 
investigation of the rocks of W. Shropshire that lasted for more than 30 years. Over 
that period, he amassed an unprecedented number of brachiopods, trilobites, 
graptolites and miscellaneous invertebrates. The brachiopod collection alone, which 
is mainly composed of moulds, includes over 3000 specimens well enough preserved 
to provide statistical information and about half as many identifiable, albeit in- 
complete, impressions. In fact the collection is not only a tribute to the patience, 
acute observation and indefatigable energies of Professor Whittard and his assistants, 
but also a measure of his brilliance as a systematist. Thus when the specimens, all 
clearly numbered, arrived in the Queen's University of Belfast in 1962, they were so 
unambiguously documented in relation to his field maps that after his untimely 
death in 1966 only a few score could not be precisely assigned to a locality, although 
the stratigraphic horizon of each was accurately known. Almost invariably, 
therefore, it has been possible to supplement Whittard's original directions and 
measurements (given in yards) for the identification of fossiliferous localities by 
citing National Grid references, which indicate position within the kilometre squares 
of the Ordnance Survey maps. Indeed only two unlocated specimens have been 
described and figured in the present systematic account, and although this has been 
done because the specimens are unique, they have not been made the types of any 
new taxa. 

The systematic study itself was begun in 1965. But for various reasons it was 
discontinued until recently, when it became evident that such a protracted delay 
in publication of a description of the collection was gravely detrimental to our 
understanding of British Ordovician palaeogeography and hence of subduction 
zones of ancient continental plates. In a regional context, the work may be regarded 
as a supplement to Whittard's own scholarly monographs on the Ordovician tri- 
lobites of W. Shropshire. Together with Dr W. T. Dean's forthcoming publication 
of a geological map of the Shelve area based on Whittard's field studies, and Dr I. 
Strachan's identification of the graptolitic assemblages, they give a much more 
comprehensive and detailed assessment of the Ordovician stratigraphy and palaeont- 
ology of the Shelve area than ever before. 

The collection was given to the writer by Professor Whittard in expectation that 
it would ultimately find its way to one of the national repositories where it could be 
adequately housed and curated. Consequently, all specimens, except for a few lost 
in the preparation of rock sections, have been donated to the Department of 
Palaeontology of the British Museum (Natural History) : their registered numbers, 
lying in the ranges BB 35305-35600 and BB 37107-37162, are given. 



ORDOVICIAN BRACHIOPODA 9 

II. FAUNAL DISTRIBUTION 

The brachiopods collected from the Ordovician rocks of the Shelve district have 
been assigned to the 85 identifiable taxa described in the systematic section, with 
19 species classified as Inarticulata, 41 as Orthida, 5 as Triplesiida, 15 as Stro- 
phomenida, 3 as Pentamerida and 1 each as Rhynchonellida and Spiriferida. The 
taxa are unevenly distributed in the succession with : 

19 species recorded from 36 localities of Mytton Flags (A), 

7 ,, ,, ,, 18 ,, „ Hope Shales (B), 

8 ,, ,, ,,12 ,, ,, Stapeley Volcanic Group (C), 
5 ,, ,, ,, 16 ,, ,, Weston Beds (D), 

8 ,, „ ,, 23 „ ,, Betton Beds (E), 

17 ,, ,, ,, 41 ,, ,, Meadowtown Beds (F), 

16 ,, ,, ,,52 ,, ,, Rorrington Beds (G), 

21 „ „ „ 8 „ „ Spy Wood Grit (H), 

14 „ ,, ,, 8 ,, ,, Aldress Shales (I), 

15 „ „ „ 4 „ „ Hagley Shales (J), 
and 23 ,, ,, ,, 4 „ ,, Whittery Shales (K). 

In the following comprehensive list of stratigraphic distribution, the lower case 
letters Y (rare), 'fc' (fairly common), 'c' (common), and 'vc' (very common) 
respectively represent the retrieval of one or more specimens of a listed species in 
0-25%, 26-50%, 51-75%, 76-100% of all localities for any one formation, while the 
capital letters 'A', 'B', 'C etc. at the head of the list are keys to formations in the 
order given above. The list is : 

ABCDEFGHI JK 

Apsotreta sp. r __________ 

Astraborthis uniplicata gen. et sp. nov. r — — — — — — — — — — 

Bicuspina modesta sp. nov. ______ rc ___ 

Bicuspina subquadrata sp. nov. — — — — — — — — rfcvc 

Bystromena perplexa gen. et sp. nov. _______ vc __ _ 

Caeroplecia plicata gen. et sp. nov. _________ rvc 

Conotreta stapeleyensis sp. nov. — r r — ______ _ 

Dalmanella elementaria sp. nov. r __________ 

Dalmanella salopiensis sp. nov. --- — rcr — - — — 
Dalmanella salopiensis gregaria sp. et 

subsp. nov. _______ vc __ 

Dalmanella salopiensis transversa sp. et 

subsp. nov. ________ f c __ 

Desmorthis ? sp. nov. r — r — — — — — — — — 

Diplarelasma sp. r __________ 

Dolerorthis cf . tenuicostata Williams __________ r 

Drabovia cf . fascicostata Havlidek _______ f c ___ 

Eocramatia dissimulata gen. et sp. nov. _ r _________ 

Euorthisina cf . moesta minor Havlicek r __________ 

Furcitella sp. __________ r 

Gelidorthis cf. partita (Barrande) ______ r ____ 



H 


I 


J 


K 


vc 


- 


- 


r 


r 


- 


— 


- 



10 SHELVE DISTRICT 

A B C D E F G 

Glyptorthis viriosa sp. nov. ----- r r 

Harknessella cf. subplicata Bancroft _______ 

Hesperonomia sp. r __________ 

Heterorthis sp. ------rfcr -- 

Horderleyella cf. plicata Bancroft _______ f cr __ 

H order leyella sp. _____ r _____ 

Kiaeromena cf. kjerulfi (Holtedahl) _________ rr 

Kjaerina (Hedstroemina) sp. _______ f c ___ 

Kullervo sp. _____ r _____ 

Lenorthis cf. proava (Salter) f c __________ 

Leptaena cf. ventricosa Williams --------- fee 

Lingulella displosa sp. nov. - — — -rrfc-r-- 
Lingulella displosa petila sp. et subsp. nov. --fc — — — — — — — — 

Mcewanella sp. __________ r 

Monobolina plumbea (Salter) c __________ 

Murinella sp. _____ r _____ 

Nicolella cf. actoniae (Sowerby) _________ rvc 

Nocturniella sp. r __________ 

Obolus subditivus sp. nov. _ c _________ 

Obolus sp. r __________ 

Onniella ostentata lepida subsp. nov. -------- fc vc vc 

Onniella sp. _ ______ r ___ 

Orbiculoidea sp. __ r ________ 

Orthambonites exopunctata sp. nov. _________ feve 

Orthambonites sp. ________ r __ 

Orthis cf. callactis Dalman r __________ 

Orthis sp. __ r ________ 

Oslogonites ?sp. r __________ 

Oxoplecia cf. nantensis MacGregor _____ r _____ 

Palaeoglossa attenuata (Sowerby) ---fccvcc c r -r 

Palaeoglossa myttonensis sp. nov. c __________ 

Palaeostrophomena sp. __________c 

Parastrophinella musculosa sp. nov. _______ f c ___ 

Parastrophinella sp. _________ r _ 

Paterula cf. bohemica Barrande rfc--r ______ 

Paterula cf. perfecta Cooper _____ r rfee -- 

Petrocrania dubia sp. nov. ________ r rfc 

Platystrophia caelata sp. nov. _________ feve 

Platystrophia cf . major Williams _______ f c ___ 

Plectorthis whitteryensis sp. nov. _________ r r 

Plectorthis sp. ________ r __ 

Protoskenidioides revelatus gen. et sp. nov. r __________ 

Pseudolingula spatula sp. nov. rr-c-fcr---- 

Rafinesquina delicata sp. nov. — — — — rfc — — — — — 

Rafinesquina sp. ______ r _ r _ r 

Reuschella horderleyensis carinata subsp. nov. — — — — — — — — — eve 

Rostricellula sparsa Williams _______ r ___ 

Salacorthis costellata gen. et sp. nov. _______ f c ___ 

Salopia cf. salteri (Davidson) _______ r ___ 

Salopia sp. _________ r fc 

Schizocrania salopiensis sp. nov. - fc r fc fc r c c - - - 

Schizotreta transversa sp. nov. _____ rr ____ 

Schizotreta sp. r r r r _______ 



ORDOVICIAN BRACHIOPODA n 

ABCDEFGHI JK 

Schmidtites ? simplex sp. nov. ____ r r f c _ f c _ _ 

Schmidtites ? simplex subcircularis sp. et 

subsp. nov. r _f c -------- 

Sericoidea cf . abdita Williams _________ r _ 

Skenidioides cf. costatus Cooper — — — — — — rfc — — c 

Sowerbyella cf. antiqua Jones _____ r _____ 

Sowerbyella multiseptata sp. nov. _______ c ___ 

Sowerbyella cf . sericea permixta Williams ________ f c __ 

Sowerbyella cf . sericea (Sowerby) --------- fcvc 

Sowerbyella sp. ______ r ____ 

Tazzarinia elongata sp. nov. _____ r _____ 

Tissintia immatura (Williams) _____f c _____ 

Tissintia prototypa (Williams) ---cfc ------ 

Triplesia sp. __________ r 

Whittardia paradoxica gen. et sp. nov. __________ c 

Zygospira sp. _______ r ___ 

The list does not include the few poorly preserved inarticulate brachiopods 
recovered from the Stiperstone Quartzite which were : a discinid, an obolid and 
Pseudolingula sp. 

The large number of identifications reflect the assiduity of the collectors and the 
great range of geologic time represented by the strata from which specimens have 
been recovered, rather than any intrinsic richness of faunas. Compared with 
contemporaneous faunas from Scotland or the Baltic, for example, the Shelve 
assemblages are restricted in content. This relative paucity may be illustrated 
(Text-fig. i) by estimating the number of species recorded per 300 m of rock for 
each formation, using thicknesses given by Watts (1925) and Whittard (1952). The 
greatest diversity of about 76 species per 300 m for the Whittery Shales is only 
one-third that calculated for the penecontemporaneous Balclatchie Mudstones 
(Williams 1962 : 27). This contrast is further emphasized by the negligible taxo- 
nomic communality, even at generic level, between the Shelve assemblages and 
coeval Scottish or Baltic ones, and appears to be related to profound environmental 
differences ranking with provincial distinctions (Williams 1973 : 246) in Recent seas. 
Accordingly only assemblages found in Wales, E. Shropshire and, to a lesser extent, 
N. Africa and Bohemia are like those of the Shelve area ; and even these show 
discrepancies in stratigraphic distributions and frequencies of occurrence probably 
linked to the benthic habits of brachiopods. 

The brachiopods occurring in the Mytton Flags have an Arenigian aspect even in 
relation to assemblages characteristic of more remote regions. Desmorthis, Dipare- 
lasma and Hesperonomia are reminiscent of American successions (Ulrich & Cooper 
1938 : 27-28) ; Orthis callactis (Rubel 1961 : 142) and Oslogonites (Opik 1939 : 118) 
of the Baltic region ; and Euorthisina and Nocturniella (Havlicek & Vanek 
1966 : 50) of Czechoslovakia, although the species with which the Shelve Euorthisina 
has been compared occurs in the Llanvirnian Sarka Formation of Bohemia and the 
Llanvirnian Tachilla Shales of Morocco (Havlicek 1971 : 75). With respect to more 
precise correlation with other Anglo- Welsh successions, only Lenorthis proava has 



12 



SHELVE DISTRICT 



80 



60 



CO 

V 50 

o 

<D 
Q. 



o~> 



40 



30 

E 20 

c 
lOH 




ABCDE FGH 

formations 

Fig. i . Graph showing the estimated number of species recorded per 300 m of : Mytton 
Flags (A) ; Hope Shales, Stapeley Shales and Weston Beds (B) ; Betton Beds (C) ; 
Meadowtown Beds (D) ; Rorrington Beds (E) ; Spy Wood Grit (F) ; Aldress Shales (G) ; 
Hagley Shales (H) ; and Whittery Shales (I). 



proved useful so far. In its type area of Anglesey, this species occurs in Arenigian 
rocks (Bates 1968 : 147) and is associated with Hesperonomiella which is probably 
closely related to the Hesperonomia of the Shelve area. L. proava also occurs in 
the contemporaneous Henllan Ashes of the Arenig district and authenticated reports 
of its widespread occurrence are in sharp contrast to the lack of information on the 
distribution of the other common Shelve brachiopod, Monobolina piumbea, or of 
inarticulates generally in the Welsh successions. Clearly these early Ordovician 
assemblages are in urgent need of investigation. 

Despite taxonomic neglect of the older Llanvirn brachiopod faunas, a fairly close 
likeness between the Llanvirn -Llandeilo assemblages of W. Shropshire and Car- 
marthenshire can be discerned. This is especially true when one takes into account 
the longer ranging constituents of the fauna like Lingulella, Palaeoglossa, Paterula, 
Pseudolingula and Schizomania which occur in both areas. However these species, 
although indicative of similar biofacies, are not very useful for correlation which is 
dependent on comparatively few articulate species. 

Tissintia prototypa is confined to the Weston and Betton formations of the Shelve 
area, both of which should now be assigned to the Upper Llanvirnian Didymograptus 
murchisoni zone according to I. Strachan (in Whittard 1966 : 297). The species is 



ORDOVICIAN BRACHIOPODA 13 

also narrowly ranging in the Llandeilo area where it occurs in ashy shales of the 
Didymograptus bifidus zone and in the succeeding basal beds of the Ffairfach Grit 
at Ffairfach. Tissintia immatura, on the other hand, is found throughout the 
Llandeilo Flags (Williams 1953 : 190-194), while its so-called 'variety' T. immatura 
plana is also recorded from conglomeratic sandstones (Williams 1953 : 185) recog- 
nized as the top member of the Ffairfach Group-specimens identified as this sub- 
species, from the basal Ffairfach Grit at Garn-wen, actually belong to T. prototypa 
(Williams 1953 : 184). In the Shelve area, T. immatura is restricted to the Meadow- 
town Beds ; presumably its absence from the Rorrington Beds, which are now 
correlated with the Middle and Upper Llandeilo Flags (Bassett in Williams et al. 
1972 : 33), may be attributed to adverse environmental conditions. The distribution 
of Tissintia in Shropshire and Carmarthenshire therefore suggests that the Betton 
Beds and at least part of the Weston Beds should be correlated with the Ffairfach 
Group, and the Meadowtown Beds with the Lower Llandeilo and, possibly, the top 
conglomeratic member of the Ffairfach Group. 

Other brachiopods tend to confirm this correlation. Sowerbyella antiqua is known 
from the Betton Beds and from the lower part of the Meadowtown Beds, and although 
it is common at certain horizons in the Ffairfach Group and Lower Llandeilo, it is 
unknown from the Middle and Upper Llandeilo Flags. The small Dalmanella 
found in both areas constitute two morphologically distinctive stocks. But the 
differences between them may have been phenotypic, and it is noteworthy that they 
first appeared in the Betton and Ffairfach successions simultaneously with Sower- 
byella. Oxoplecia nantensis, which was first described by MacGregor (1961 : 196) 
from the Upper Llandeilo rocks of the Berwyn Dome, is widely distributed in the 
Llandeilo strata of S.W. Wales and a related form has been recovered from the 
Ffairfach Group. Horderleyella on the other hand, which is well represented in 
Welsh successions, rarely occurs, and then only as an indeterminate species in the 
Meadowtown Beds. Moreover the Shropshire strophomenid has surprisingly 
proved to be a new species of Rafinesquina unrelated to its well-known contemporary 
Macrocoelia llandeiloensis. This unexpected discovery suggests that the Welsh 
Llanvirn-Llandeilo strophomenids require revision because those previously identi- 
fied as the species 'llandeiloensis' may prove to belong to two different genera. 

The graptolite-bearing Rorrington Beds and the trilobite-rich Middle and Upper 
Llandeilo Flags have different biofacies. Each has yielded a subordinate but 
distinctive brachiopod assemblage, with a predominance of inarticulates in the 
Rorrington Beds and of Tissintia and Dalmanella in the Llandeilo Flags. Indeed 
articulate brachiopods found in the Rorrington Beds are usually immature specimens 
indeterminate at infrageneric level, although Bicuspina, Heterorthis and Skenidioides, 
as precursory elements of the Spy Wood fauna, deserve attention. The first two, 
together with Gelidorthis cf . partita, are reminiscent of Middle Caradocian assemblages 
of Morocco (Havlicek 1971) and Czechoslovakia (Havlicek and Vanek 1966 : 54-55). 
This earlier appearance in the Anglo-Welsh successions than in Bohemian or African 
strata of articulate brachiopods seems to have been part of a recurrent pattern, 
possibly representing the effects of diachronic migration . Thus Tissintia is not known 
in Morocco until mid-Llandeilo times ; and Tazzarinia, which has been collected from 



i 4 SHELVE DISTRICT 

Meadowtown Beds, first occurs in Morocco in the Lower Caradocian part of the 
Ktaoua Formation. 

The greater diversity of post-Rorrington brachiopod faunas affords an opportunity 
to apply Bancroft's Caradocian stages (1945) to the younger Shelve formations and 
thereby effect a correlation with the classic sections of E. Shropshire. 

The Spy Wood Grit, with Harknessella cf. subplicata, Horderleyella cf. plicata, 
Salopia cf. salteri and young specimens of Heterorthis, may be correlated with the 
Coston Beds (Dean 1958 : 218). The latter also contain undescribed Bicuspina, 
small Dalmanella, Glyptorthis, and Sowerbyella, which may prove to be conspecific 
with those found in the Shelve area. Further afield in Wales, the additional pre- 
sence of Platystrophia cf. major, Skenidioides cf. costatus and Salacorthis suggests 
contemporaneity with the early Caradocian faunas of the Arenig district and 
Anglesey (Whittington and Williams 1955 ; Bates 1968), and W. Carmarthenshire 
and Pembrokeshire (R. Addison pers. comm.). 

The Aldress Shales are less decisively dateable because the pre-Soudleyan bra- 
chiopods of E. Shropshire are largely undescribed. But the occurrence of Sower- 
byella cf . sericea permixta is an indirect link through its first appearance within the 
Nant Hir Mudstones of the Bala district (Bassett, Whittington & Williams 
1966 : 263-264) and suggests that the formation is at least Harnagian in age. 

The relative richness of the Hagley and Whittery faunas is in keeping with their 
identification as part of an evolving Nicolella association (Williams 1973 : 242). 
This association is known to have occupied the Anglo-Welsh Basin sporadically 
throughout Caradocian times, being prevalent in the Costonian Derfel Limestone and 
Longvillian Gelli-grin Calcareous Ashes of the Bala district (Williams 1963 : 341) 
and the Upper Longvillian to Actonian formations inclusive of E. Shropshire (Dean 
1958 : 221-224). Consequently the Shelve assemblages appear to be partly an 
anachronistic blend of earlier and later forms. Thus Caeroplecia, Palaeostrophomena, 
Platystrophia and Salopia are also known from the Derfel Limestone although they 
are represented by different species, but another distinctive member of the associa- 
tion, Dolerorthis tenuicostata, is not. In contrast to these similarities, Nicolella 
actoniae (or a subspecies), Onniella ostentata and Sowerbyella sericea are known from 
Longvillian successions in N. Wales and/or E. Shropshire. Finally Bicuspina and 
Reuschella are unknown from pre-Soudleyan strata and Kiaeromena cf. kjerulfi and 
Leptaena ventricosa first appear in the Soudleyan Allt Ddu Mudstones. Indeed, 
bearing in mind the extraordinary range of N. actoniae including its occurrence in the 
Cautleyan Portrane Limestone (Wright 1964 : 165), and the complex relationship 
among the Sowerbyellas of N. Wales which are mainly closely related to S. sericea, 
the Hagley and Whittery brachiopod assemblages are most likely to be Soudleyan 
in age. 

Like the older brachiopod assemblages, the Caradocian faunas of Shelve contain 
very few taxa of use in establishing even a tentative correlation with the biostrati- 
graphic successions of other provinces. Only the Caradocian rocks of Bohemia and 
Morocco have yielded faunas with an overall similarity, although the majority of 
common elements again first appear in demonstrably younger rocks than in Shrop- 
shire in Bohemia (Havlicek and Vanek 1966 : 53-57). Heterorthis, Horderleyella, 



ORDOVICIAN BRACHIOPODA 



15 



Onniella and Reuschella, for example, first occur in the Zahorany Formation, 
Dalmanella in the Letna Formation and Bicuspina and Drabovia in the Liben Forma- 
tion, which have been dated as Soudleyan, Harnagian and Costonian respectively 
(Williams in Williams et al. 1972 : fig. 2). Moreover, even so characteristic a 



idard Succession 


Shelve Succession 
Whittery Shales 


Soudleyan 
CARADOC 


Whittery Volcanic Gp. 


Hagley Shales 


Hagley Volcanic Gp. 


Harnagian 


Aldress Shales 


Costonian 


Spy Wood Grit 


Upper 

Middle 
1 1 Awncn n 


Rorrington Beds 


LLMINUtlHJ 

Lower 


Meadowtown Beds 


Upper 


Betton Beds 


Weston Beds 


LLANVIRN 


Stapeley Shales 


Lower 


Stapeley Volcanic Gp. 




Hope Shales 


ARENIG 


Mytton Flags (s.l.) 


Stiperstones Quartzite 



Fig. 2. Classification of the Ordovician rocks of the Shelve area according to the 

brachiopod faunas. 



16 SHELVE DISTRICT 

Bohemian stock as Drabovia is represented in the Costonian Spy Wood Grit by D. 
fascicostata, which is known in Czechoslovakia only from the allegedly younger 
Letna Formation. 

In summary then, it appears that although the W. Shropshire assemblages of 
Ordovician brachiopod species are strongly endemic and not reliable indices for 
correlation outside the Anglo-Welsh Basin, they do confirm the time-stratigraphic 
classification of the Shelve succession proposed by Dean (in Whittard 1967 : 317) 
except for the equation of the Rorrington Beds with the Middle and Upper Llan- 
deilian rather than the basal Caradocian (Text-fig. 2). 



III. FAUNAL ASSOCIATIONS 

The calculation of density distributions of species solely according to the frequency 
of their presence in localities for any given formation, as in the preceding section, is 
effective enough for determining the degree to which a taxon is diagnostic of a 
particular stratigraphic horizon. Such estimates, however, do not give any indica- 
tion of the relative commonness of occurrence of species recovered from any single 
locality : of how, for example, Palaeoglossa myttonensis and Protoskenidioides 
revelata are recorded from 13 and 3 localities of Mytton Flags respectively, although 
the former species is represented in the collection by only 16 pedicle or brachial 
valves compared with 25 valves of the latter. Such data may be very important. 
They may reflect differences in the transporting capacity of sea currents. Alter- 
natively, the Palaeoglossa populations may have consisted of thinly spread individuals 
tolerant of widely differing environments whereas Protoskenidioides may have existed 
as high density clusters of shells in rigorously circumscribed ecological niches. 

Different kinds of data are required to promote investigations into the palaeo- 
environment on the one hand and palaeoecosystems on the other. Palaeoenviron- 
mental studies require information on the number of each kind of valve as well as 
the proportion of complete to fragmented shells composing collections. This 
demand is well served by data prepared in the manner adopted by Temple (1968 : 9 ; 
1970 : 8) for Lower Llandovery brachiopods. Yet for this approach, information 
on the size range frequencies of species is equally important ; and since such details 
cannot be derived from normal statistical tables, they are rarely available and the 
exercise correspondingly diminished in its efficacy. In contrast, studies of relation- 
ships within and between communities mainly require estimates of species distribu- 
tion among communities ; frequencies of individuals representing a species within 
an association only indicate its importance according to numerical ranking. Never- 
theless, such data are valuable from both palaeoecologic and stratigraphic view- 
points. Variation in the numbers of individuals present in fossil communities or 
their residues may be aids in determining the optimum conditions for the main- 
tenance of a species in space, or in identifying changes in the interrelationship between 
species of an evolving association in time. In his study of the Shelve trilobites, 
Whittard (1966 : 298) illustrated variation in the number of specimens collected from 
each formation by time-histograms. Brachiopods pose a different problem because 



ORDOVICIAN BRACHIOPODA 



17 










01 


CO 




60 





nJ 


hH 






h 




c 


00 





l-t 


+■> 




-(-• 




>, 


r^ 


§ 


M 



O d 



o 








T3 


M 




V 






M 






0) 


M 




> 


M 


















<D 







Ih 


M 




U) 






4J 






'O 


Ov 


M 


0) 




W 

m 


en 


CO 


s3 


en 


r^ 




+■> 







SH 






O 






CO 


in 




V 






j> 












13 

> 


■*■ 










cd 






IS 


tn 




u 






ni 






)h 






-O 


N 



w I 



I H M M I 

M I I 
I I I 
H N I 

ro <N 1 

I M I 
M 

M M I 

I IO I 

I H | 

I H I 

M I I 

(N t-^ I M 

IO H ( H 

tN I I 

I M I 

I rO I 

I <N I 

M <N I 

M t-t | 

C^ M M 

I ro I 

I <o I 



I I I 



M M I 



T3 

<ti 



£ 

(3 
CO 

H 



•^ X 



s 
. s 

co ■» 



Oh 



ft 

« <■> • 






f hi 



13 •« « 
<s o «> 



5^ 9 



•S .§ 

© S 

o S 

S» ■*» 

?» Li 






•3 



o ^ 
O O 



11 

. ^» <» 

M § o 

°" <s . 

■a § ° 

S &c-S 

© S S 

c> «S <a 

*C£ *-^ ■** 

<o « « 

0,0, 



a -a 

<» s 

*■ « 

•is "> 

•« © 

5 s 

© <» 

0,0, 






* « S 



■Hi 
Co to 



18 SHELVE DISTRICT 

each individual is, initially at least, represented in a death assemblage by two valves 
so that counts of all specimens in collections consisting mainly of disarticulated valves 
could be a gross misrepresentation of frequencies in vivo. Consequently the number 
of individuals of listed species collected from each locality, shown in Tables i to II, 
are the totals of, the more commonly occurring valves whether they are brachial or 
pedicle. Also, the tables do not include localities yielding specimens of only one 
species on the grounds that their distribution does not contribute to our under- 
standing of the relationship within brachiopod associations. Such data may, 
however, assist in identifying those species most tolerant of palaeoenvironmental 
variation and have, therefore, been referred to in the explanatory text when this 
possibility appears to hold. 

Table i shows the frequency distributions of various brachiopod species collected 
from 22 localities of Mytton Flags. The fauna was dominated by an association 
including Monobolina, Palaeoglossa and Lenorthis (each also occurring alone in 5, 5 
and 2 localities respectively) which was especially prevalent in unlaminated micaceous 
siltstones and fine-grained crystal tuffs. Another association of Dalmanella, 
Protoskenidioides and Euorthisina was characteristic of laminated siltstones and 
shales. The Monobolina association has a few elements in common with an associa- 
tion involving Obolus, Paterula and Pseudolingula found in the dark grey micaceous 
Hope Shales (Table 2). A similar lithofacies within the Stapeley Volcanic Group 
(Table 3) supported another inarticulate association with Lingulella, Schmidtites and 
Schizocrania, although Orthis and Desmorthis, which may have been constituents of 
an association descended from the Arenig Dalmanella one, occur in crystal-lithic 
tuffs. 

Table 2 

The number of pedicle or brachial valves of the listed species recovered 
from 7 localities of Hope Shales 

1234567 

Eocramatia dissimulata I — — I 4 — — 

Obolus subditivus -31 1 - 6 1 

Paterula cf. bohemica 1 1 1 - - 1 2 

Pseudolingula spatula - 1 - 1 - - 1 

Schizocrania salopiensis - 1 — - 1 1 - 

Table 3 

The number of pedicle or brachial valves of the listed species recovered 
from 5 localities of Stapeley Shales 

12345 

Conotreta stapeleyensis 4 — - - — 

Desmorthis ? sp. — 1 - 2 — 

Lingulella displosa petila — — 1 — 1 

Orthis sp. - 3 — - - 

Schizocrania salopiensis — — II — 

Schizotreta sp. - - - 1 - 

Schmidtites ? simplex subcircularis 1 - - - 1 



ORDOVICIAN BRACHIOPODA 19 

The succeeding Weston and Betton Beds (Tables 4, 5) consist mainly of laminated 
siltstones. They sustained an association which had evolved out of the Arenigian 
Monobolina complex and was evidently ancestral to the Meadowtown assemblage. 
In the Weston Beds, the association is dominated by Pseudolingula, Palaeoglossa 
and Tissintia with the last occurring on its own in five localities additional to those 
set out in Table 4. By Betton times, changes in the relative importance of species 
had taken place so that the association was made up mainly of Palaeoglossa, Tissintia 
and Schizomania with each of the last two stocks constituting the only brachiopods 
found in three extra localities. 



Table 4 

The number of pedicle or brachial valves of the listed species recovered from 
10 localities of Weston Beds 

123456789 10 

Palaeoglossa attenuata I I I — — i — — — i 

Pseudolingula spatula 1421 1211-12 

Schizocrania salopiensis 1--1 3 — — - — - 

Schizotreta ? sp. ____ 2 — — — — — 

Tissintia prototypa 14 - - - - - 115 3 4 

Table 5 

The number of pedicle or brachial valves of the listed species recovered from 11 localities of 

Betton Beds 

1 23456789 10 11 

Dalmanella salopiensis _________5_ 

Lingulella displosa _____ j _____ 

Palaeoglossa attenuata 2151111-1-1 

Paterula cf. bohemica _________ 1 _ 

Rafinesquina delicata _________6- 

Schizocrania salopiensis 1--1 1--1 1 - - 

Tissintia prototypa - 1 20 - — - 7 2 — — 5 

The Meadowtown Beds (Table 6) are essentially laminated calcareous siltstones or 
fine-grained sandstones with subordinate calcarenites. This lithofacies was occupied 
by an association derived from the later Llanvirn ones but incorporating new ele- 
ments, so that the dominant forms were Palaeoglossa, Dalmanella, Rafinesquina, 
Tissintia and Pseudolingula. Palaeoglossa was the only brachiopod occurring in 
nine additional localities. The genus must have been more tolerant of variation in 
the palaeoenvironment than its contemporaries because it is also the dominant 
member (with Schmidtites) of an inarticulate association, descended from the early 
Llanvirn communities, which includes Lingula, Paterula and Pseudolingula and is 
found in rarely occurring medium grey shales and mudstones. 

The subordinate association identified in the Meadowtown Beds became over- 
whelmingly important in the Rorrington lithofacies which consisted of medium to 
dark grey slightly calcareous shales. The principal community (Table 7) was 



SHELVE DISTRICT 



•s 

pq 
a 

o 



o 








CD 




> 




o 




o 




o 




Ih 


o 


09 






w 


o 
u 


1-1 


a 


n 


to 


< 


-3 


H 


4) 



CO 

O M I 



S W 



u M 



I 



00 

<N 



c~» m co 

<N N 



M 


H 


1 


<N 


1 


1 




tr, 


1 


CO 
<N 


M 


m 



<N CO IN 



M ^ 



o 



o 

4) 

a. 



M <N 

00 | 

IH 

»• I 

M 

O M 

M 

m I 

CO CO 

M 

(N CO 
H 

H t*- 

M 

O M 

IH 

O 00 
00 o 
O I 

* I 



I 

<S 

co 

M po 
<N 



M I 



(O I I INI 

I I I M | 

* I I I I 

I I I I I 

en I I I M 

lO I M I M 

n io aoo I 

N M pi O M 

I 



M CN I 

I N I 

H O | O 1 

M | | 

I I I 

I I I 

I M | 

I I I 

I CN I 

N CO I 

en t^ I 

1 M I 

M 

0) ■* M 



Hill 



I I I en I I 



I I I 

I H I 

I <s I 

m en I 

in in I 

M 



I 

M | 

(N I 

M I 

I I 

I I 

I I 

I I 

I I 

I I 

I I 

I I 

M- I 



I I 

I I 

I I 

I I 

I I 



I 
I 
I 
en 

I 
I 

CO 



00 



I 

o 

M 

I 

I 

I 
I 
I 
I 

o 

I 

CO 



in 

CO 



W 
iJ 

m 
< 
H 







X 


<M 


T3 

3 


(N 


o 


1- 
(N 


a 


CO 
M 



« M 



On m 



8 00 

O H 

<N M 

S *° 

o H 

£i <n 

(U M 

% J? 

o M 



o 

CD 

0) 

1) 



a 






O o 



J3 
o 



<u 

a, 



1 M 


1 


M 




1 


| | H H 


1 




<N 


1 M 


1 


M 




1 


1 CN 


1 


1 




IH 


M M 


1 


IH 




N 


Tf CO 


M 


N 




«T 


1 N 


M 


1 




1 


1 M 
1 1 


1 
1 


M 

N 




IH 

M 


M M 


1 


1 




M 


<N vO 


1 


1 




1 


M CN 


1 


1 


M 


H 


1 1 


1 


H 


IH 


M 


1 N 


IH 


IH 




■<i- 


•n o 

M 


M 


M 
M 




t^ 


1 M 


1 


in 




1 


1 M 


1 


M 




1 


M | 


1 


o 




1 


<N CO 


M 


00 




00 


M CO 


H 


H 




1 


1 M 


in 


N 


M 


1 


1 M 


1 


M 




1 


1 01 


1 


O 
IH 




1 


1 *-* 


1 



M 




1 


1 IH 
1 1 


M 
1 


1 




1 

1 


M M 


1 


1 




1 


1 !>■ 


1 


1 




*■ 


CO O 


1 


N 




* 



<D 



3 
C 
CD 

Xi 

H 



$." 



P, 



» -S 



3s ".3 



s 
« 
s 






O S 






S « J 



« . 

r 

S 



« o 






s « 

«3 ^ 






■» « « 



QO^^ 



H <3 5 






it 



* 5 
8 a 



« -5 

1.9 



s 

cC ft; to <A co co 



. ^; i^; .s; Ss 
u u u ,o 



51 5 



6 

a 

H 



8 



« S3 
8 S 

|i 

3 J 
•8 « 

oq Q 



Si 



■2 e 
s | 
S -ft. 



CO ^ 

| 8 



<o Ch' 



8 « 

i ^ 

<S S 

ts Si 

ft, ft. 



- to 

8 

<o « 



■9,8 



a) 
« 8 



^ 



u Dh 
en 



8mm 

(Kk "^* *^k 



w « ^ 



„ g i 

ft! en co co co co 



ORDOVICIAN BRACHIOPODA 21 

composed of Palaeoglossa, Schmidtites and Schizocrania with the last genus alone 
occurring in six extra localities and each of the first two in an additional five. 
Immature shells of Rafinesquina are most widely distributed, followed by Dalmanella 
and Glyptorthis. These presumably represent unsuccessful waves of colonization by 
an association akin to those characteristic of the Meadowtown Beds and the succeed- 
ing calcareous Spy Wood Grit (Table 8). The Spy Wood community is dominated 

Table 8 

The number of pedicle or brachial valves of the listed species recovered 
from 7 localities of Spy Wood Grit 





i 


2 


3 


4 


5 


6 


7 


Bicuspina modesta 


- 


- 


4 


2 


i 


50 


12 


Bystromena perplexa 


i 


2 


2 


- 


i 


8 


7 


Dalmanella salopiensis gregaria 


- 


3 


7 


2 


i 


76 


25 


Drabovia cf . fascicostata 


i 


- 


i 


- 


- 


12 


- 


Glyptorthis viriosa 


i 


- 


i 


I 


i 


21 


4 


H arknessella cf . subplicata 














1 


Heterorthis sp. 


- 


- 


- 


- 


i 


10 


2 


Horderleyella cf . plicata 


- 


- 


- 


- 


- 


3 


1 


Kjaerina (Hedstroemina) sp. 


- 


- 


- 


- 


- 


1 


1 


Onniella sp. 


- 


- 


- 


I 


- 


- 


- 


Palaeoglossa attenuata 


i 


I 


- 


- 


- 


18 


4 


Parastrophinella musculosa 


- 


- 


- 


- 


- 


6 


2 


Paterula cf . perfecta 


- 


3 


- 


I 


- 


4 


- 


Platystrophia cf. major 


- 


- 


- 


- 


- 


1 


1 


Rostricellula sparsa 


- 


- 


- 


- 


- 


5 


- 


Salacorthis costellata 


- 


- 


i 


- 


- 


5 


2 


Salopia cf . salteri 


- 


- 


- 


- 


- 


2 


- 


Schizocrania salopiensis 


i 


- 


- 


I 


i 


13 


3 


Skenidioides cf. costatus 


- 


- 


4 


- 


- 


3 


- 


Sowerbyella multiseptata 


- 


i 




I 


- 


24 


11 


Zygospira sp. 


- 


- 


- 


- 


- 


1 


- 



by articulate brachiopods like Dalmanella, Glyptorthis, Bystromena, Bicuspina and 
Sowerbyella, although Palaeoglossa and Schizocrania persist. 

The Aldress Shales, like many of the older formations, bear traces of two distinct 
associations (Table 9). The more commonly occurring is typified by Paterula and 
Schmidtites which have also been recovered from three and one extra localities 
respectively, and is found in medium grey micaceous siltstones. Crystal tuffs, on 
the other hand, supported a relatively prolific fauna dominated by Sowerbyella, 
Dalmanella and Onniella. 

The fossil-bearing members of the two youngest Ordovician formations preserved 
in the Shelve area, the Hagley and Whittery Shales, consist mainly of light olive- 
grey calcareous tuffs and mudstones. However, the Whittery mudstones contain 
lithic and crystal fragments of volcanic debris up to 1 cm in size, mud balls and 
broken shells, which suggest that even those parts of the formation not obviously 
derived from volcanic ash falls may have been deposited by lahars. Indeed the 
only non-volcanic sediment appears to be a light olive-grey micaceous siltstone 



22 shelve district 

Table 9 

The number of pedicle or brachial valves of the listed 
species recovered from 3 localities of Aldress Shales 

1 2 3 

Bicuspina subquadrata - - 1 

Dalmanella salopiensis transversa - 1 33 

Heterorthis sp. - - 1 

Horderleyella sp. - - 7 

Onniella ostentata lepida - 2 22 

Orthambonites sp. - - 1 

Palaeoglossa attenuata 1 - - 

Paterula cf. perfecta 1 - - 

Petrocrania dubia - — 1 

Platystrophia cf. major - - 1 

Plectorthis sp. - - 1 

Rafinesquina sp. - - 1 

Schmidtites ? simplex 1 - - 

Sowerbyella sericea permixta - 2 66 

occurring in the Hagley Shales. This lithofacies bears Sericoidea, while the rest of 
the formation contains an association descended from that occurring in the Aldress 
tuffs but with Onniella and Reuschella dominant, Sowerbyella playing a subordinate 
role and Dalmanella no longer occurring (Table 10). This fauna attained its full 

Table 10 

The number of pedicle or brachial valves of the listed species 
recovered from 4 localities of Hagley Shales 

1234 

Bicuspina subquadrata - - 1 1 

Caeroplecia plicata - - - 1 

Kiaeromena sp. 2 - - - 

Leptaena cf. ventricosa 2 1 - - 

Nicolella cf. actoniae - - - 1 

Onniella ostentata lepida 4 1 1 3 

Orthambonites sp. 1 - - 1 

Parastrophinella sp. - - - 1 

Petrocrania dubia - - - 1 

Platystrophia caelata 1 - - 1 

Plectorthis whitteryensis - - - 2 

Reuschella horderleyensis carinata 1 - 2 1 

Salopia sp. - - - 1 

Sericoidea cf. abdita - 10 - - 

Sowerbyella cf. sericea 2 - - 2 

diversity in the Whittery mudstones where Bicuspina, Caeroplecia, Nicolella, 
Onniella, Orthambonites, Reuschella and Whittardia are very common (Table 11). 
Many constituents, including Caeroplecia, Nicolella, Orthambonites, Palaeostro- 
phomena and Platystrophia, reveal that the fauna is mainly composed of the Nicolella 



ordovician brachiopoda 23 

Table ii 

The number of pedicle or brachial valves of the listed species 
recovered from 4 localities of Whittery Shales 

1234 

Bicuspina subquadrata 2 1 17 15 

Caeroplecia plicata 3 - 14 7 

Dolerorthis cf. tenuicostata - - 3 - 

Furcitella sp. — — 1 — 

Glyptorthis viriosa — — 2 — 

Kiaeromena sp. 1 — — — 

Leptaena cf. ventricosa - 1 1 1 

Mcewanella sp. - - - 2 

Nicolella actoniae 3 9 6 5 

Onniella ostentata lepida 4 3 9 25 

Orthambonites exopunctata 4 7 12 11 

Palaeoglossa attenuata 1 — — - 

Palaeostrophomena sp. 1 1 1 - 

Petrocrania dubia - - 1 3 

Platystrophia caelata 1222 

Plectorthis whitteryensis - - 7 2 

Rafinesquina sp. — - 1 - 

Reuschella horderleyensis carinata 12 11 6 3 

Salopia sp. — - 2 1 

Skenidioides cf. costatus 1 - 2 1 

Sowerbyella cf. sericea 2 1 1 8 

Triplesia sp. - - - 1 

Whittardia paradoxica 3-82 

association (Williams 1973 : 242), but elements like Bicuspina and Leptaena 
further suggest that remnants of the Dinorthis association may also be present. 

In conclusion, it is noteworthy that the composition of brachiopod assemblages 
collected from the Ordovician succession of Shelve appears to be closely connected 
with the nature of the entombing sediments. In general, two main groups seem to 
be involved : an inarticulate association connected with the finer clastic sediment 
and a predominantly articulate one sustained in the coarser, banded, sporadically 
calcareous siltstones or fine sandstones. Both associations may be traced throughout 
much of the Ordovician succession with one dominant over the other. Both show 
profound changes in the composition and relative importance of their constituent 
taxa and, thereby, greatly reduce the prospects for precise interprovincial correlation. 



IV. SYSTEMATIC METHODS 

The procedure adopted during the systematic study of the Shelve brachiopods was 
to quantify, wherever possible, any morphological variation displayed by conspecific 
specimens from any one locality or horizon and evaluate differences detected among 
congeneric samples by standard statistical tests for significance. This method is 
similar to that used by the writer during researches on Girvan and Bala faunas 
(Williams 1962 : 69-79 > I 963 : 333), and requires only a few words of explanation. 



24 



SHELVE DISTRICT 



The number of individuals attributable to any species represented in the collection 
is small enough to warrant not only the invariable use of small sample techniques but, 
for many taxa, nothing more than the mean (or mode) and range of variability. 
Continuous univariate characters which are assumed to be normally distributed, 
like the distances at which folds originate anterior of umbones or the wavelengths of 
ribs at a given distance antero-medially of the umbo, have been compared by the 
Mest. Discontinuous univariate characters, like the number of costae ornamenting 
valves or the relative branching of costellae enumerated according to Bancroft's 
notation (Bancroft 1945 : 186 ; Williams 1962 : 77), have been compared by x 2 tests, 
or contingency or two-by-two tables dependent on the size of samples. 

Continuous variables derived by measuring the shell or its several parts, like the 
cardinalia, dental plates or muscle impressions, are expressions of incremental 
proportional growth, which may or may not be allometric, and all show very high 
positive (or rarely negative) correlations with one another. The significance of any 
differences in estimates of such parameters has been determined by statistical analysis 
of a series of bivariate distributions. The method is adequate for the comparison 
of such closely inter-related estimates of shell growth. It also has the advantage of 
not obscuring the identity of those characters that underwent changes in the rate of 
growth during cladogenetic or phylogenetic speciation. The development of sig- 
nificantly longer dental plates in one of a number of compared species, for example, 
will be apparent whether the lengths of plates are consistently paired with the 
lengths, maximum widths or depths of the pedicle valves in compared samples. 

In estimating variation in outline, shape and relative size of the brachiopod 
skeleton, very many measurements in almost any plane or direction may be taken, 
provided the same disposition is adopted in measuring all individuals in compared 
samples. However, certain orientations for the measurement of shells or their 
impressions have become widely favoured by brachiopod researchers over the years 





Fig. 3. Diagrammatic representation of moulds (and adherent shell) of Orthambonites to 
show the orientation adopted for measuring certain dimensions including : length (1), 
maximum width (w) and thickness (th) of a valve, the wavelength of a costa (wv), the 
length of dental plates (dl), the length (Isc) and width (wsc) of muscle scars, the length of 
a median ridge (Is), and the length (lc) and width (wc) of brachiophore bases. 



ORDOVICIAN BRACHIOPODA 25 

and may be regarded as standard. These include length (sagittal), width (maxi- 
mum) and depth (maximum) . Certain orientations for the measurement of internal 
features are also coming into general use although they have not yet been standard- 
ized. Those used in this paper are illustrated in Text-fig. 3, in relation to internal 
moulds which lend themselves very well to estimating the forward extensions of the 
bases of dental plates, brachiophores and sockets, or of muscle scars. 

V. SYSTEMATIC PALAEONTOLOGY 

Class INARTICULATA Huxley 1869 

Order LINGULIDA Waagen 1885 

Superfamily LINGULACEA Menke 1828 

Family OBOLIDAE King 1846 

Subfamily OBOLINAE King 1846 

OBOLUS Eichwald 1829 

Obolus subditivus sp. nov. 

(PI. 1, figs. 1-8) 

Diagnosis. Biconvex, subcircular Obolus almost as long as wide with a lenticular 
ventral pseudointerarea extending anteriorly for 15% of the length of the pedicle 
valve and a dorsal pseudointerarea not divided into propareas by an identifiable 
median groove. 

Description. Subcircular, biconvex, Obolus with a mean length relative to 
width of 95% (range 83% to 111%) and a mean depth relative to length of 14% 
(range 12% to 17%) for 6 pedicle valves, and a brachial valve about one-tenth as 
deep as long, both valves evenly convex transversely but flattened slightly anteriorly ; 
beaks marginal ; shell surface ornamented by concentric growth lines and up to 10 
impersistent low narrow plicae. 

Striated pseudointerarea of pedicle valve lenticular in outline, orthocline to ana- 
cline in attitude, extending forward for 15% of the length of 6 pedicle valves (range 
12% to 20%) and divided into 2 propareas by a shallow pedicle groove bearing a 
fine low median ridge ; ventral muscle impressions unknown. 

Striated pseudointerarea of brachial valve obtusely triangular, orthocline, extend- 
ing forward for only one-twentieth of the length of the brachial valve, median groove 
not identifiable ; a pair of elongate muscle scars, possibly representing lateral muscle 
bases, are vaguely impressed on either side of a low obscure median ridge. 

Type material. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 35573a, b) 5-0 4-6 

Paratypes External and internal moulds of pedicle valve 

(BB 35574a, b) 3-3 3-i 

External and internal moulds of brachial valve 

(BB 35575a, b) 3*8 3-8 



26 SHELVE DISTRICT 

Horizon and locality. Hope Shales exposed in path west of Brithdir Farm, 
i mile ENE of Old Church Stoke (Grid Ref. SO 300953). 

Discussion. The attribution of relatively small obolids found in the Hope Shales 
to Obolus s.s. is a matter of convenience until more is known about the musculature 
of the new species, which was not accommodated by well-developed platforms as in 
more typical species. Even the identification of the species as an obolinid may be a 
matter of dispute because the dorsal pseudointerarea does not bear an identifiable 
median groove although it is certainly a better developed feature than the striated 
thickening of the posterior margin characteristic of the glossellinid brachial valve. 
In respect of both features, the new species differs from orthodox Obolus. However, 
Cooper (1956 : 189-193) assigned a few species to the genus which bear some 
resemblance to 0. subditivus. This is especially true of 0. ? biconvexa Cooper from 
the Lower Paper ville Formation of Tennessee and the closely related 0. ? nitens 
Cooper from the Lower Rich Valley Formation of Virginia, although both are more 
elliptically elongate and lack radial plicae, while the former is further distinguished 
by the presence of concentric undulations. 

Obolus sp. 

(PI. 1, figs. 9, 10) 

In contrast to 0. subditivus sp. nov., a few exfoliated valves have been recovered 
from the Mytton Flags, which are more representative of typical Obolus. The ex- 
posures in Bergam Quarry above Tankerville Mine (Grid Ref. SO 355995) have 
yielded an evenly convex pedicle valve (BB 35581a, b) about 6 mm long, which is 
transversely oval in outline with a length about 70% of the width and a depth 25% 
of the length. A heart-shaped postero-medial portion of the valve has broken away 
and probably constitutes the muscle platform. No identifiable impressions are 
preserved on the internal mould of the platform but a well-defined pedicle groove is 
seen medially indenting a relatively narrow pseudointerarea. 

Another incomplete specimen (BB 35582) about 8 mm wide and over 1 mm deep 
has been collected from near an adit 290 yds. west of Wood House (Grid Ref. 
SJ 336002). It bears a pair of submedial muscle scars and may be the brachial 
valve of the same species. Even so, paucity of material precludes specific identifica- 
tion. 

SCHMIDTITES Schuchert & Le Vene 1929 

Schmidtites ? simplex sp. nov. 
(PL 1, figs. 11-15) 

Diagnosis. Subequally biconvex, circular obolids with valves almost as long as 
wide and 10% as deep as long, ornamented by fila and overlapping lamellae ; 
pseudointerareas narrow, probably striated and divided by relatively wide pedicle 
groove and median dorsal depression. 

Description. Subequally biconvex, circular obolids with inconspicuous beaks and 
valves 99% as long as wide and 10% as deep as long ; ornamented by strong fila, 



ORDOVICIAN BRACHIOPODA 27 

becoming less distinct medially, and finely developed overlapping lamellae ; however 
the thin shell is usually preserved in a collapsed state commonly expressed as trans- 
verse or even radiating wrinkles. 

Ventral interior with a narrow crescentic striated pseudointerarea divided by a 
relatively wide pedicle groove ; muscle impressions known only in one specimen as a 
pair of suboval, submedial scars, possibly flanked by another set, distinguishable as 
smooth areas in a finely pitted internal surface. 

Dorsal pseudointerarea with a very narrow thickened zone along the posterior 
margin which may represent a medially grooved pseudointerarea ; other internal 
features unknown although a low median ridge extends forwards for three-quarters 
the length of some valves. 

Type material. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 37134a, b) 4-5 4-5 

Paratypes External and internal moulds of brachial valve 

(BB 37135a, b) 

Exterior of pedicle valve (BB 37136) 

External and internal moulds of pedicle valve 

(BB 37137a, b) 

Internal mould of brachial valve (BB 37138) 

Internal mould of pedicle valve (BB 37139) 

External and internal moulds of pedicle valve 

(BB 37140a, b) 3-5 3-5 

Horizon and localities. Rorrington Beds : BB 37134, 37138 from exposures 
in Holywell Brook 160 yds WNW of the Mount, Rorrington (Grid Ref. SJ 299007) ; 
BB 37135, 37139 from outcrops in Deadman's Dingle 100 ft upstream from junction 
with Spy Wood Brook (Grid Ref. SO 280960) ; BB 37136 from exposures 60 yds 
north-west of the Mount, Rorrington (Grid Ref. SJ 302005) ; BB 37137, 37140 from 
the tributary to Lower Wood Brook 340 yds south-east of Desert (Grid Ref. 
SJ 308017). 

Schmidtites ? simplex subcircularis sp. et subsp. nov. 
(PL i, figs. 16, 17 ; PI. 2, fig. 1) 

Diagnosis. Like Schmidtites ? simplex but subcircular with valves 96% as wide 
as long and 7% as deep as long ; protegulum semicircular about 0-2 mm long. 

Type material. 

length width (mm) 
Holotype Exterior of pedicle valve (BB 37141) 2-5 2-8 

Paratypes External and internal moulds of pedicle valve 

(BB 37142a, b) 3-0 3-0 

External and internal moulds of brachial valve 

(BB 37143a, b) 3-3 3-6 



3-o 


27 


3-5 


4-0 


2-6 


— 


4-2 


4-5 


4-0 


- 



28 SHELVE DISTRICT 

Horizon and locality. All specimens from Stapeley Shales exposed in Whits- 
burn Dingle 650 yds WSW of Lord's Stone, Leagh (Grid Ref. SJ 328019). 

Discussion. Subcircular inarticulates, occurring fairly abundantly in the Stape- 
ley Shales and the Meadowtown and Rorrington Beds, are provisionally assigned to 
the obolid genus Schmidtites only because of the lack of muscle platforms which are 
normally elaborate structures in other obolinids. With regard to the modest differ- 
entiation of the posterior margins, here interpreted as poorly developed pseudo- 
interareas, the new species differs even from other described Schmidtites, and may 
eventually prove to be generically separable from all other lingulaceans, to which 
superfamily it almost certainly belongs. 

Notwithstanding the obscurity of internal features and simplicity of shape, the 
conspecific Meadowtown and Rorrington samples are significantly different 
(p < 0-02) from specimens recovered from the Stapeley Shales which became pro- 
gressively longer than wide during growth (Table 12). The difference has been 
recognized by the erection of a new subspecies. 

Table 12 

Statistics of length (1) and maximum depth (w) of n pedicle valves of Schmidtites ? simplex sp. nov. 

from the Rorrington Beds (A) and the Meadowtown Beds (B), and of Schmidtites ? simplex 

subcircularis sp. et subsp. nov. from the Stapeley Shales (C) 





A 


B 


C 


n 


29 


20 


9 


1 mm (var 1) 


2-88 (1-148) 


2-98 (1-235) 


3-13 (3-23) 


w mm (var w) 


2-9(1-145) 


3-01 (1-244) 


3-00 (1-952) 


r 


o-995 


0-977 


0-986 


logel (var logel) 


0-9927 (0-1296) 


1-0288 (0-1296) 


o-i (0-2844) 


log e w (var log e w) 


0-9996 (0-1279) 


1-0358 (0-129) 


1-0004 (°' I 9°3) 


r e 


o-995 


0-979 


099 


a (var a) 


0-9931 (0-00037) 


Q-9975 (0-00231) 


0-8307 (0-00195) 



Subfamily LINGULELLINAE Schuchert 1893 

LINGULELLA Salter 1866 

Lingulella displosa sp. nov. 
(PI. 2, figs. 2-8) 

Diagnosis. Ventribiconvex, elongately oval to subtriangular Lingulella with a 
brachial valve 87% as wide as long and a pedicle valve up to 13% as deep as long, 
ornamented by impersistent concentrically arranged lamellae, pedicle groove and 
dorsal pseudointerarea extending anteriorly for 9% and 7% to the valve length 
respectively. 

Description. Ventribiconvex Lingulella normally changing in outline from 
elongately oval to subtriangular with a broadly rounded to truncated anterior margin 
by an acceleration in antero-lateral shell accretion during adult growth stages ; 
brachial valve 87% as wide and less than 10% as deep as long, pedicle valve about 
13% as deep as long and 87% as wide as long (range 77% to 97% for 6 specimens), 



ORDOVICIAN BRACHIOPODA 29 

both valves evenly convex in longitudinal and transverse profiles ; external surface 
ornamented by concentrically disposed impersistent lamellae giving a finely scalloped 
appearance ; ventral beak acute, dorsal beak obtusely rounded. 

Ventral pseudointerarea striated, slightly anacline to orthocline, crescentic in 
outline and divided into 2 curved propareas by a shallow, relatively wide pedicle 
groove extending forward for 9% of the valve length ; umbonal muscle scars suboval 
located immediately anteriorly of the propareas ; 2 short, low ridges narrowly diverge 
from the anterior end of the pedicle groove. 

Dorsal pseudointerarea striated, orthocline, lenticular in outline and extending 
forward for 7% of the length of the brachial valve ; dorsal muscle impressions 
unknown ; postero-medial internal surfaces of both valves sporadically and 
shallowly pitted. 

Type material. 

length width (mm) 

Holotype Incomplete external and internal moulds of 

pedicle valve (BB 37144a, b) 14-0 11-5 

Paratypes External and internal moulds of pedicle valve 

(BB 37146a, b) 12-0 io-o 

External and internal moulds of brachial valve 

(BB 37147^, b) 15-0 ii-o 

External and internal moulds of pedicle valve 

(BB 37148a, b) 14-0 12-0 

Internal mould of brachial valve (BB 37149) 14-5 14-0 

Horizons and localities. Meadowtown Beds : BB 37144 from exposures 
80 yds from Meadowtown Quarry along the Rorrington road (Grid Ref. SJ 311012) ; 
BB 37146 from a small excavation in the corner of a field alongside the cart-track 
from Meadowtown to Waitchley 330 yds due south of Waitchley (Grid Ref. 
SJ 311017) ; BB 37147 from outcrops 570 yds south-west of Meadowtown Quarry 
(Grid Ref. SJ 306009) ; BB 37148 from outcrops along the lane to Lower Ridge from 
bench mark 754, Little Weston (Grid Ref. SO 293984) ; BB 37149 from Rorrington 
Beds exposed in Lower Wood Brook 400 ft south-east of the Meadowtown -Rorring- 
ton road (Grid Ref. SJ 307008). 

Lingulella displosa petila sp. et subsp. nov. 
(PI. 2, figs. 9-11 ; PL 3, fig. 1) 

Diagnosis. Gently biconvex, suboval Lingulella with a brachial valve 70% as 
wide as long, ornamented by impersistent concentrically arranged lamellae, pedicle 
groove and dorsal pseudointerarea extending anteriorly for 13% and 8% of the valve 
length respectively. 

Description. Subequally biconvex Lingulella normally suboval in outline with 
well-rounded anterior margins and a maximum width at about 60% of the shell 
length anterior of the umbones ; pedicle valve about 5% as deep as long and, on 
average, 70% as wide as long (range 63% to 80% for 5 specimens), both valves 



30 SHELVE DISTRICT 

gently and evenly convex in longitudinal and transverse profiles ; external surface 
ornamented by concentrically disposed impersistent lamellae giving a finely scalloped 
appearance ; ventral beak sharply acute, dorsal beak less conspicuous and more 
rounded. 

Ventral pseudointerarea striated, anacline to orthocline, crescentic in outline and 
divided into two obliquely disposed propareas by a shallow pedicle groove extending 
forward for about 13% of the valve length ; ventral muscle impressions unknown. 

Dorsal pseudointerarea striated, orthocline, lenticular in outline and extending 
forward for 5% of the valve length ; dorsal muscle impressions unknown ; fine 
median ridge faintly developed for over half the valve length, postero-medial internal 
surfaces of both valves sporadically and shallowly pitted. 

Type material. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 37150a, b) io-o 6-0 

Paratypes External and internal moulds of brachial valve 

(BB 37151a, b) 8-o 6-o 

External and internal moulds of pedicle valve 

(BB 37152a, b) - 6-o 

Type horizon and locality. All specimens from the Stapeley Volcanic Group 
exposed in Whitsburn Dingle, 650 yds WS W of Lord's Stone, Leagh (Grid Ref . 
SJ 328019). 

Discussion. Ordovician Lingulella from the Shelve area appear to differ from 
other known Lingulella in being ornamented by finely scalloped lamellae. The pat- 
tern is reminiscent of Glyptoglossella although in other respects, notably the presence 
of a dorsal pseudointerarea, the specimens cannot even be assigned to the same sub- 
family. The change in outline from oval to subtriangular that took place during the 

Table 13 

Statististics of length (1) and maximum width (w) of n brachial valves of Lingulella displosa sp. 
nov. from the Rorrington Beds (A) and the Meadowtown Beds (B) 

A B 

n 12 8 

1 mm (var 1) 8-84 (14-383) 8-05 (25-209) 

w mm (var w) 8-6 (13-647) 6-97 (19-902) 

r 0-986 0-97 

a (var a) 0-9741 (0-00269) 0-8885 (°' 00 774) 

Table 14 

Statistics of length (1) and maximum depth (w) of 8 pedicle valves of Lingulella displosa sp. nov. 

from the Rorrington Beds 

1 mm (var 1) 4-76 (8-931) 

w mm (var w) 4-37 (8-494) 

r 0-994 

a (var a) 0-9752 (0-00184) 



ORDOVICIAN BRACHIOPODA 31 

growth of L. displosa is also unusual, although a similar acceleration in growth along 
antero-lateral vectors was responsible for the subtriangular appearance of adult 
shells of the Tremadocian L. lingulaeformis (Mickwitz) from the Baltic area (Gory- 
anski 1969 : 38), the mid-Ordovician L. decorticata Cooper from the Elway Formation 
of Virginia, and L. rideanensis Cooper from the Aylmer Formation of Ontario 
(Cooper 1956 : 198, 203). All three species, however, are described as having an 
ornamentation consisting solely of fine concentric lines. 

Lingulella collected from the Meadowtown and Rorrington Beds are indistinguish- 
able in every respect (Tables 13, 14) and constitute L. displosa s.s. The small 
sample derived from the Stapeley Shales, on the other hand, differs significantly in the 
relative narrowness and very gentle biconvexity of the shell and in the trace of a 
fine median septum in the brachial valve. These differences have prompted the 
erection of a new taxon, but to indicate its essential affinity with the Meadowtown 
and Rorrington forms this is at the subspecific level only. 



PALAEOGLOSSA Cockerell 1911 emended A. W. 

1848 Glossina Phillips : 370 ; non Wiedemann 1830. 

191 1 Palaeoglossa Cockerell : 96. 

1956 Palaeoglossa Cockerell ; Cooper : 194. 

1965 Palaeoglossa Cockerell ; Rowell in Williams et al. : H270. 

Diagnosis. Elongately oval, biconvex lingulellinids ornamented by fine dis- 
continuous fila with the dorsal pseudointerarea not differentiated into propareas. 

Description. Elongately oval, biconvex lingulellinids with an acute curved 
ventral beak ; surface ornamented by fine fila in addition to growth lines ; ventral 
pseudointerarea slightly anacline divided by a narrow pedicle groove into striated 
propareas with strong flexure lines, dorsal pseudointerarea orthocline, striated, 
lenticular in outline and not divided into propareas ; low median ridge sporadically 
developed in the brachial valve ; muscle impressions unknown except for traces of a 
pair of umbonal muscle scars immediately anterior of the pseudointerareas of both 
valves. 

Type species. Lingula attenuata J. de C. Sowerby in Murchison 1839. 

Type horizon. Meadowtown Beds. 

Discussion. Well-preserved valves from the Meadowtown Beds afford an 
opportunity to learn more about the internal morphology of the topotypic Lingula 
attenuata Sowerby which is the type species of Palaeoglossa Cockerell (now. subst. pro 
Glossina Phillips 1848). The presence and degree of differentiation of the pseudo- 
interareas indicate that the genus is best accommodated among the Lingulellinae. 
Indeed in the absence of information about the principal muscle scars, Palaeoglossa 
differs from Lingulella only in details. They include the greater biconvexity, the 
more incurved beaks and the more elongately oval outline of Palaeoglossa, and 
possibly the skeletal structure, because the Palaeoglossa shell tends to be more laminar 
and its inner layers are not radially striated as in Lingulella. Differences between 
Palaeoglossa and Pseudobolus Cooper (1956 : 194) as currently understood are even 



32 SHELVE DISTRICT 

more finely drawn ; and when the internal features of the latter genus become 
known, it may prove to be a junior synonym of Palaeoglossa. 

Palaeoglossa attenuata (Sowerby) emended A. W. 
(PI. 3, figs. 2-13) 

1839 Lingula attenuata Sowerby in Murchison : 641. 
1866 Lingula attenuata Sowerby ; Davidson : 44. 

Diagnosis. Subequally biconvex Palaeoglossa with a brachial valve 73% as 
wide and 6% as deep as long and a pedicle valve 71% as wide and 7% as deep as 
long ; fila fine and discontinuous ; ventral pseudointerarea with a shallow narrow 
pedicle groove extending anteriorly for 11% of the length of the pedicle valve, and a 
dorsal pseudointerarea and low median ridge extending forward for 7% and 43% 
of the brachial valve length respectively. 

Description. Subequally biconvex, elongately oval Palaeoglossa with curved 
lateral and rounded anterior margins, brachial valve 73% as wide as long and 6% as 
deep as long (with a range of 3% to 8% for 9 valves), pedicle valve 71% as wide as 
long and 7% as deep as long, both valves evenly convex in longitudinal and trans- 
verse profiles ; external surface ornamented by growth lines and fine discontinuous 
fila about 0-25 mm in wavelength ; ventral beak acutely incurved, dorsal beak 
obtusely rounded. 

Ventral pseudointerarea striated, slightly anacline, crescentic in outline, 11% as 
long medially as the length of the valve, pedicle trough narrow and shallow ; divided 
umbonal muscle scars rarely impressed. 

Dorsal pseudointerarea striated, almost orthocline, lenticular in outline and 
extending forward medially for 7% of the length of the valve (range 6% to 8% for 6 
specimens) ; divided umbonal muscle scars rarely impressed on either side of a low 
variably developed median ridge extending anteriorly for 43% of the length of the 
valve (range 28% to 66% for 5 specimens). 

Figured material. 

External and internal moulds of brachial valve (BB 37109a, b) 

External and internal moulds of pedicle valve (BB 37110a, b) 

External and internal moulds of pedicle valve (BB 37111a, b) 

Exterior of pedicle valve (BB 37112) 

Exterior of pedicle valve (BB 37113) 

External and internal moulds of brachial valve (BB 37114a, b) 

Internal mould of pedicle valve (BB 37115) 

External and internal moulds of pedicle valve (BB 37116a, b) 

External and internal moulds of disarticulated valves 

(BB 37117a, b) 

External and internal moulds of brachial valve (BB 37118a, b) 

Horizons and localities. Meadowtown Beds : BB 37111, 37114 from outcrops 
on the Meadowtown -Rorrington road, 220 yds from Meadowtown Chapel (Grid Ref. 



sngth 


width (mm) 


11-0 


8-0 


7-0 


5-0 


12-5 


87 


8-3 


6-o 


6-5 


4-5 


13-0 


9-0 


16-0 


- 


5-5 


3-6 


55 


3-o 


12-0 


9-0 



ORDOVICIAN BRACHIOPODA 33 

SJ 309012) ; BB 37113, 371 15 from outcrops 80 yds from Meadowtown Quarry 
along Rorrington road (Grid Ref. SJ 311013) ; BB 37110 from outcrops along lane 
to Lower Ridge from benchmark 754, Little . Weston (Grid Ref. SO 293984) ; 
BB 37109 from outcrops in Lower Wood Brook 400 ft south-east of the Meadow- 
town -Rorrington road (Grid Ref. SJ 307008) ; BB 371 12 from laneside exposures 
370yds north of Meadowtown Chapel (Grid Ref. SJ311015). Betton Beds: 
BB 37116, 37117 from outcrops downstream of junction between Holywell and 
Whitehouse Brooks (Grid Ref. SJ 303004) ; exact locality of BB 371 16 uncertain. 
Rorrington Beds : BB 37118 from exposures in a stream 530 yds north-west of 
Meadowtown Chapel (Grid Ref. SJ 308017). 



Palaeoglossa myttonensis sp. nov. 
(PI. 4) figs. 1-5) 

Diagnosis. Ventribiconvex Palaeoglossa with a brachial valve 78% as wide and 
4% as deep as long and a pedicle valve about twice as deep ; fila fine and normally 
continuous ; dorsal pseudointerarea lenticular and restricted to the umbonal area, 
dorsal median septum extending anteriorly for about half the valve length ; postero- 
medial internal surfaces of shell sporadically and shallowly pitted. 

Description. Ventribiconvex, oval Palaeoglossa with curved lateral and rounded 
anterior margins ; brachial valve 78% as wide and 4% as deep as long, pedicle 
valve about 8% as deep as long ; both valves evenly convex in longitudinal and 
transverse profiles ; surfaces ornamented by growth lines and fine, somewhat 
distant, more or less continuous fila with a wavelength of about 0-25 mm. 

Ventral pseudointerarea striated, slightly anacline, crescentic, divided by a 
shallow pedicle trough extending anteriorly for 11% of the length of the valve. 

Dorsal pseudointerarea striated, slightly anacline, lenticular in outline, restricted 
to the umbonal area and extending anteriorly for 9% of the valve length ; dorsal 
median ridge low, extending forward for 51% of the valve length (range 47% to 
54% for 5 specimens) ; postero-medial internal surfaces of both valves sporadically 
and shallowly pitted. 

Type material. 

length width (mm) 
Holotype External and internal moulds of conjoined 

valves (BB 37119) - 4-5 

Paratypes Incomplete external and internal moulds of 

brachial valve (BB 37121a, b) - - 

External and internal moulds of slightly 
disarticulated valves (BB 37122a, b) 6-5 5-0 

Internal mould of brachial valve (BB 37123) 6-5 5-2 

Horizon and localities. Mytton Flags : BB 37119, 37122 from tip heaps on 
road to Squilver 150 yds ENE of Whitegrit School (Grid Ref. SO 323976) ; BB 37121 
from sides of adit of New Perkin's Level (Grid Ref. SJ 376022) ; BB 37123 from 



34 SHELVE DISTRICT 

outcrops near the top of steep tributary to Crowsnest Dingle, 350 yds WNW from 
Blakemoorflat (Grid Ref. SJ 373008). 

Discussion. Representatives of Palaeoglossa have been recovered from most 
formations constituting the Ordovician successions of the Shelve area, but samples 
giving some indication of the variability in the shape of the shell and the develop- 
ment of internal features have been obtained from only the Mytton Flags and the 

Table 15 

Statistics of length (1) and maximum width (w) of n brachial valves of Palaeoglossa attenuata 

(J. de C. Sowerby) from Meadowtown Beds (A) and the Rorrington Beds (B) and of Palaeoglossa 

myttonensis sp. nov. from the Mytton Flags (C) 





A 


B 


C 


n 

1 mm (var 1) 

w mm (var w) 

r 
a (var a) 


29 
9-31 (11-789) 
6-82 (6-243) 
0-951 
0-7277 (0-00186) 


5-01 (16-285) 

3'53 (7-37) 

0-994 

0-6727 (0-00045) 


10 

5-42 (2-813) 

4-19 (i-93) 

0-981 

0-8283 (0-00321) 




Table 


16 





Statistics of length (1) and maximum width (w) of n pedicle valves of Palaeoglossa attenuata 
(J. de C. Sowerby) from the Meadowtown Beds (A), the Rorrington Beds (B) and the Betton 

Beds (C) 

ABC 

n 32 36 23 

1 mm (var 1) 10-76 (28-081) 5-8 (11-584) 5-09 (1-589) 

wmm(varw) 7-67 (17-835) 4-23(6-518) 3-23 (0-67) 

r 0-982 0-984 0-918 

logel ( var log e l) 2-2667(0-2173) 1-6098(0-2958) 1-5978(0-0594) 

log e w (var log e w) 1-9051(0-2646) 1-2872(0-3104) 1-1431(0-0618) 

r e 0-985 0-9871 0-925 

a (var a) 1-1036(0-0012) 1-0244(0-00079) 1-0203(0-00718) 

Table 17 

Statistics of length (1) and maximum depth (th) of 9 pedicle valves of Palaeoglossa attenuata 
(J. de C. Sowerby) from the Meadowtown Beds 

1 mm (var 1) 10-34 (9*89) 
th mm (var th) 0-77 (0-087) 

r 0-964 

a (var a) 0-094 (0-00008) 

Table 18 

Statistics of length (1) and length of ventral pseudointerarea (pi) of 7 pedicle valves oi Palaeoglossa 
attenuata (J. de C. Sowerby) from the Meadowtown Beds 

1 mm (var 1) 13-5 (66-oo) 
pi mm (var pi) 1-43 (0-789) 

r 0-909 

a (var a) 0-1093 (0-00041) 



ORDOVICIAN BRACHIOPODA 35 

Betton, Meadowtown and Rorrington Beds. The Meadowtown assemblage is 
topotypic with those specimens on which J. de C. Sowerby (in Murchison 1839 : 
pi. xxii, fig. 13) based P. attenuata. Comparisons of internal features like the relative 
anterior extension of the pseudointerareas and the dorsal median ridge, as well as 
the depth and outline of the valves (Tables 15-18), show that those Palaeoglossa 
recovered from the Betton and Rorrington Beds are indistinguishable from P. 
attenuata. 

The small sample taken from the Mytton Flags, on the other hand, differs from the 
others in the significantly greater expansion in width relative to length during shell 
growth (Table 15). The beak of the brachial valve is also more acutely rounded and 
the lenticular pseudointerarea correspondingly more restricted to the umbonal 
region, while the sporadically distributed shallow pits indenting the postero-medial 
internal surfaces of the Mytton shells have not been seen in younger specimens. 
These differences merit specific recognition ; and it is interesting to note that the new 
species appears to be restricted in stratigraphic distribution because the few specimens 
of Palaeoglossa known from the Weston, Spy Wood and Whittery Formations are 
best assigned to attenuata. 



Subfamily GLOSSELLINAE Cooper 1956 
PSEUDOLINGULA Mickwitz 1909 emended A. W. 

1909 Pseudolingula Mickwitz : 771. 

1945 Pseudolingula Mickwitz ; Sinclair : 58. 

1965 Pseudolingula Mickwitz ; Rowell in Williams et al. : H267. 

1969 Pseudolingula Mickwitz ; Goryanski : 41. 

Diagnosis. Biconvex, parallel-sided glossellinids ornamented by fila with an 
orthocline ventral pseudointerarea divided by a pedicle groove into two striated 
propareas and an undifferentiated dorsal posterior margin ; umbonal muscle scars 
divided, other muscle bases, including identifiable lateral and transmedian sets, 
supported on a broad adnate platform in the pedicle valve and impressed on either 
side of a median septum in the brachial valve ; interiors sporadically pitted postero- 
medially. 

Discussion. Well-preserved internal moulds of Pseudolingula recovered from 
the Ordovician rocks of the Shelve area reveal more details of the internal mor- 
phology of this genus than have hitherto been recorded, and allow for a reappraisal 
of its position within the lingulide hierarchy. All characters confirm the lingulacean 
affinities of the taxon, and the divided aspect of the umbonal muscle bases indicates 
that it is best assigned to the Obolidae. This allocation is corroborated by the 
development of a pedicle groove bounded by flexures and striated propareas at least 
in the specimens from the Rorrington Beds. The undifferentiated dorsal posterior 
margin, however, suggests that Pseudolingula is closer to the Glossellinae than the 
Lingulellinae as currently conceived (see Rowell in Williams et al. 1965 : H266-269), 
and its transference to the former subfamily is accordingly adopted for this study. 



36 



SHELVE DISTRICT 



Pseudolingula spatula sp. nov. 
(PI. 4, figs. 6-14 ; PL 5, fig. 1 ; Text-fig. 4) 

Diagnosis. Subequally biconvex, subquadrate Pseudolingula with a brachial 
valve 67% as wide and 5% as deep as long, ornamented by strong fila laterally, 
becoming finer medially, and bearing an adnate ventral muscle platform and a strong 
dorsal median ridge arising up to one-fifth anteriorly of the beak and extending 
forwards 58% of the length of the valve. 

Description. Subequally biconvex, subquadrate Pseudolingula with parallel 
lateral and obtusely rounded anterior margins and obtuse beaks with the more 
rounded dorsal one subtending an angle of about ioo° ; brachial valve 67% as wide 
and about 5% as deep as long ; both valves subcarinate postero-medially but 
flattening laterally and anteriorly ; external surface ornamented by growth lines 
and fila which are strongly developed with a wavelength of about 0-2 mm laterally 
but tending to become finer medially. 

Ventral interior with median depression to accommodate pedicle but nature of the 
propareas unknown ; umbonal muscles divided and represented by elongately oval 
scars on either side of the median depression, other muscle bases not identifiable 
individually but supported on a broad medially pointed platform, heavily rutted by 
growth lines and extending forward of the beak for about two-fifths of the length of 
the pedicle valves. 

Dorsal interior with median depression at the beak but posterior margin un- 
differentiated ; median septum arising up to one-fifth of the valve length anterior of 
the beak and becoming strong and relatively high at the anterior end which is, on 
average, 58% of the valve length forward of the dorsal beak (range 54% to 61% for 




pedicle groove 



umbonal scar 



muscle scar 



median septum 




A B 

Fig. 4. Diagrammatic views of (A) the ventral and (B) the dorsal interiors of 

Pseudolingula. 



ORDOVICIAN BRACHIOPODA 37 

3 specimens) ; muscle impressions including a pair of suboval umbonal scars situated 
submedially and a pair of long lateral and transmedian muscle scars inserted on 
either side of the median septum ; the two sets of scars extend anteriorly of the beak 
for 20% and 43% of the valve length respectively ; postero-medial internal surfaces 
of both valves sporadically and coarsely pitted. 

Type material. 

length width (mm) 

Holotype Incomplete external and internal moulds of 

brachial valve (BB 37124) 23-0 16 -o 

Paratypes External and internal moulds of brachial 

valve (BB 37125a, b) 10-5 6-5 

Internal mould of pedicle valve (BB 37126) 14-0 7-6 

External and internal moulds of brachial 
valve (BB 37128a, b) 13-5 9-0 

Figured Incomplete exfoliated exterior of brachial 

specimen valve (BB 37129) - 7-6 

Type horizon and localities. Weston Beds : BB 37124 from exposures in 
Cwm Dingle 640 yds south of Little Weston (Grid Ref. SO 294978) ; all other 
specimens from outcrops 210 yds south-east of road crossing Betton Dingle, Lyde 
(Grid Ref. SJ317015). 

Discussion. Pseudolingula is known to occur sporadically throughout much of 
the Ordovician succession of the Shelve area, but only the Weston Beds have 
yielded a sufficient number of specimens to allow some estimates to be made of 
variability in the shape and internal structure of the shell. The valves known from 
other horizons must also be provisionally referred to the same species as the Weston 
material. Indeed an internal mould of a pedicle valve (BB 37130a, b) from the 
Rorrington Beds exposed in Deadman's Dingle, 90 yds from the junction with Spy 
Wood Dingle (Grid Ref. SO 289958), affords the only clear evidence of the nature of 
the ventral posterior margin (PI. 4, fig. n), while an incomplete ventral interior 
(BB 37129) from the Meadowtown Beds exposed at the side of Minicop Farm 
(Grid Ref. SJ 314018), and a badly exfoliated valve up to 15 mm long (BB 35583) 
from the Mytton Flags exposed in the River Camlad (Grid Ref. SO 320919), verify 
the stratigraphic persistence of the characteristic ventral muscle platform and con- 
centric ornamentation (PI. 4, figs. 12-14). 

In the absence of quantitative data on shell outline and proportions in described 
Pseudolingula, the new species can only be distinguished, for the time being, by the 

Table 19 

Statistics of length (1) and maximum depth (w) of 18 brachial valves of Pseudolingula spatula 

sp. nov. from the Weston Beds 

1 mm (var 1) 11-84 (39' 1 6) 
w mm (var w) 7-93 (16775) 

r 0-987 

a (var a) 0-6545 (0-00067) 



38 SHELVE DISTRICT 

strong anterior development of the dorsal median septum and the strength and 
anterior configuration of the ventral muscle platform. In these aspects, the new 
species is certainly different from the type species, the mid-Caradocian Pseudolingula 
quadrata (Eichwald) from the U.S.S.R. (Goryanski 1969 : 41), as well as various 
Caradocian American forms (Sinclair 1945 : 59-62 ; Cooper 1956 : 215-217), and 
other differences may become apparent in due course. 

Family ELKANIIDAE Walcott & Schuchert 1908 

MONOBOLINA Salter 1865 

Monobolina plumbea (Salter) 

(PI. 5, figs. 2-10 ; PI. 6, fig. 1) 

1859 Lingula plumbea Salter in Murchison : 50, foss. 8, fig. 1. 

1865 Monobolina plumbea (Salter) Salter in Murchison : 334, pi. xiB, fig. 10. 

Diagnosis. Subelliptical Monobolina ornamented by fila and costellae numbering 
about 5 and 8 per mm respectively, 5 mm antero-medially of the umbo, and with well- 
developed muscle fields extending anteriorly for about half the valve length. 

Description. Subelliptical, biconvex Monobolina with an evenly convex brachial 
valve averaging 78% as long as wide and 13% as deep as long and a slightly less 
convex pedicle valve tending to become flattened anteriorly in adult shells ; shell 
surface except for protegular areas ornamented by thin sharp fila and costellae in 
densities, 5 mm antero-medially of the umbones, of 3, 4, 5 and 6 fila per mm in 1, 2, 
7 and 1 brachial valves, and of 7, 8 and 9 costellae per mm in 4, 7 and 2 brachial 
valves ; ventral pseudointerarea roughly lamellose, orthocline, averaging 72% of 
the width of the valve and 11% as long as the valve medially where it is divided by a 
strong pedicle groove 52% as wide as long, dorsal pseudointerarea poorly developed ; 
dorsal interior with a pair of variably impressed outside lateral and transmedial 
muscle scars situated submedially 44% of the valve length anterior of the dorsal 
umbo to form the postero-lateral boundaries to a pair of central muscle scars im- 
pressed medially at about the middle of the valve, vascula lateralia poorly defined ; 
ventral muscle platform low and solid with a convex boundary culminating in a sub- 
dued median ridge extending anteriorly for 46% of the valve length ; three pairs of 
scars representing the transmedian, middle lateral and anterior lateral muscles 
occurring on either side of the median ridge and flanked by a pair of anterior lateral 
impressions, vascula lateralia commonly well defined. 

Figured material. 

Internal mould of pedicle valve (BB 35467) 

External mould of pedicle valve (BB 35468) 

External and internal moulds of brachial valve (BB 35469a, b) 

External and internal moulds of brachial valve (BB 35470a, b) 

Internal mould of pedicle valve (BB 35471) 

Internal mould of pedicle valve (BB 35472) 

External and internal moulds of pedicle valve (BB 35473a, b) 



length 


width (mm) 


15-5 


20-4 


5-5 


7-5 


II'O 


- 


7-3 


10-0 


12-0 


15-8 


I7-0 


18-3 


12-5 


- 



ORDOVICIAN BRACHIOPODA 39 

Horizon and localities. Mytton Flags : BB 35472 from adit entrance in 
Maddox's Coppice, 1200 yds ENE of St Luke's Church, Snailbeach (Grid Ref . 
SJ 382030) ; BB 35468 from sides of adit running into hillside at New Perkin's 
Level (Grid Ref. SJ 376022) ; BB 35469 from runnel just east of footpath to Blake- 
moorflat at head of Mytton Batch (Grid Ref. SJ 373006), BB 35470 from exposures 
in sunken cart track, west side of Snailbeach Coppice (Grid Ref. SJ 375026) ; 
BB 35471 from near top of steep tributary to Crowsnest Dingle 350 yds WNW of 
Blakemoorflat (Grid Ref. SJ 373008) ; BB 35467 from dump below Old Perkin's 
Level, Snailbeach (Grid Ref. SJ 376022) ; BB 35473 from dump outside adit in 
Perkin's Beach just below col separating Perkin's Beach and Mytton Batch (Grid 
Ref. SJ 366002). 

Discussion. The type species of Monobolina has been well known for over a 
century as an easily distinguishable inarticulate with a restricted stratigraphic and 
geographic range. Salter (1859 : 56) based the species on specimens collected from 
Mytton Flags exposed west of the Stiperstones. Most of the specimens used in 
preparing this emended description are topotypic with his, and the remainder come 
from localities in the vicinity. They therefore afford an opportunity to present some 
reliable data on the morphologic variability of the species which illustrate the homo- 
geneity as well as the distinctiveness of the taxon. 

Table 20 

Statistics of length (1) and maximum width (w) of 19 brachial valves of Monobolina plumbea 

(Salter) 

1 mm (var 1) 8-82 (20-311) 

w mm (var w) 11-29 (3°'383) 

r 0-979 

a (var a) 1-2232 (0-00362) 

Table 21 

Statistics of length (1) and thickness (th) of 18 brachial valves of Monobolina plumbea (Salter) 

1 mm (var 1) i°'37 (16-106) 

th mm (var th) 1-39 (0-487) 

r 0-927 

log e l (var log e l) 2-2688 (0-1398) 

logeth (var log e th) 0-216 (0-2251) 

r e 0-934 

a (var a) 1-269 (0-0128) 

Table 22 

Statistics of length (1) and length of the lateral-transmedial muscle scars (sc) of 19 brachial valves 

of Monobolina plumbea (Salter) 

1 mm (var 1) H'57 (9*2) 
sc mm (var sc) 5-16 (2-163) 

r 0-903 
a (var a) 0-4849 (0-00255) 



4 o SHELVE DISTRICT 

Table 23 

Statistics of valve length (1) and length of the entire muscle fields (Is) in 12 brachial valves of 

Monobolina plumbea (Salter) 

1 mm (var 1) 12-89 (4"979) 

Is mm (var Is) 6-69 (2-343) 

r 0-943 

logel ( var log e l) 2-5419 (0-0296) 

logels (var logels) 1-8754 (0-0510) 

r e 0943 

a (var a) 1-3126 (0-0189) 

Table 24 

Statistics of length (1) and thickness (th) of 12 pedicle valves of Monobolina plumbea (Salter) 

1 mm (var 1) 8-27 (17-045) 

£R mm (var th) i-o (0-306) 

r 0-95 

logel (var logel) 2-001 (0-2228) 
logeth (var logeth) —0-015(0-0301) 

r e 0-972 

a (var a) 0-3676 (0-00075) 

Table 25 

Statistics of valve length (1) and muscle platform length (sc) of 13 pedicle valves of Monobolina 

plumbea (Salter) 

1 mm (var 1) 12-74 ( I2- 778) 
sc mm (var sc) 5-88 (3-235) 

r 0-983 

a (var a) 0-5032 (0.00079) 

Table 26 

Statistics of valve length (1) and pedicle groove length (pe) of 13 pedicle valves of Monobolina 

plumbea (Salter) 

1 mm (var 1) 12-74 ( I2 '778) 
pe mm (var pe) 1-42 (0-288) 

r 0-903 

a (var a) 0-1502 (0-00035) 

Family PATERULIDAE Cooper 1956 

PATERULA Barrande 1879 

Paterula cf. bohemica Barrande 
(PI. 6, figs. 2-11) 
1879 Paterula bohemica Barrande : plate 152. 

Diagnosis. Elongately oval Paterula with a brachial valve 87% as wide as long 
and 17% as deep as long ; ventral beak submarginal, dorsal beak submarginal, 



ORDOVICIAN BRACHIOPODA 41 

9% of the length of the brachial valve forward of the truncated posterior margin ; 
limbus well developed. 

Description. Dorsibiconvex, elongately oval Paterula with a slightly truncated 
posterior margin and a rounded anterior one, brachial valve with a mean width 
relative to length of 87% (range 82% to 90% for 7 valves) and a mean depth relative 
to length of 17% (range 15% to 20% for 5 valves), pedicle valve with a mean depth 
relative to length of 8% for 2 valves ; surface ornamented by fine concentric lines ; 
pedicle notch wide, immediately posterior of ventral beak ; submarginal dorsal 
beak located, on average, 9% of the length of 3 brachial valves (range 7% to 10%) 
forward of their posterior margins and not constituting the greatest depth of the 
valve which is in mid-region. 

Limbus well defined in both valves about 7% as wide as the valves are long ; 
other internal features unknown except for two lines diverging from the ventral 
beak and bounding a faint median ridge. 

Figured material. 

length width (mm) 

Exterior of pedicle valve (BB 35590) i-6 1-5 

Exterior of brachial valve (BB 35591) 1-8 i-6 

Exterior of pedicle valve (BB 35592) i-6 1-5 

Exterior of pedicle valve (BB 35593) i-6 1-4 

Exterior of brachial valve (BB 35594) 1*9 1-9 

Incomplete exterior of brachial valve (BB 35595) - 1-9 

External and internal moulds of pedicle valve (BB 35588a, b) 2-0 1-9 

External and internal moulds of brachial valve (BB 35589a, b) i-6 1-4 

Horizons and localities. BB 35590 to 35595 inclusive from Meadowtown 
Beds exposed 100 yds south-east of Minicop farm (Grid Ref. SJ 315018) ; BB 35588, 
35589 from Hope Shales exposed in Hope Dingle behind Hope Cottage (Grid Ref. 
SJ 338009). 



Paterula cf. perfecta Cooper 
(PL 6, figs. 12, 13) 
1956 Paterula perfecta Cooper : 258. 

Diagnosis. Elongately oval Paterula with a brachial valve 88% as wide as long 
and 20% as deep as long, ventral beak submarginal, dorsal beak 23% of the length of 
the brachial valve forward of the rounded posterior margin ; limbus becoming 
subdued anteriorly. 

Description. Dorsibiconvex, elongately oval Paterula with rounded posterior 
and anterior margins, brachial valve with a mean width relative to length of 88% 
(range 80% to 90%) and a mean depth relative to length of 20% (range 20% to 21%) 
for 4 valves ; pedicle valve with a mean depth relative to length of 11% (range 10% 
to 11%) for 3 valves ; surface ornamented by fine concentric lines ; pedicle notch 
small, ventral beak submarginal, located on average 10% of the length of 3 pedicle 



42 SHELVE DISTRICT 

valves (range 9% to 11%) forward of their posterior margins ; dorsal beak represent- 
ing the highest point of the brachial valve located, on average, 23% of the length of 
5 brachial valves (range 15% to 25%) forward of their posterior margins. 

Limbus well defined in both valves except posteriorly where it becomes obscure, 
other internal features unknown except for the characteristic two lines diverging 
from the ventral beak. 

Figured material. 

length width (mm) 

Exfoliated exterior of pedicle valve (BB 35596) 2-0 i-8 

Exfoliated exterior of brachial valve (BB 35597) 20 i-8 

Horizon and locality. Spy Wood Grit exposed in Spy Wood Dingle (Grid 
Ref. SO 281959). 

Discussion. Paterula occurs sporadically throughout much of the Shelve 
succession, although only 3 samples were available for study and even they were too 
small for anything but the simplest statistical assessment of variability. They do 
show, however, that the Shropshire Paterula belong to two distinct species. In the 
older stock, as represented by the samples from the Hope Shales and Meadowtown 
Beds, the posterior margin tends to be flattened or truncated, the limbus is well 
developed and the dorsal beak is submarginal. Such features, with the exception 
of the first, are also characteristic of P. bohemica, figured by Barrande (1879 : plate 
152) from the Llanvirnian Sarka Formation of Czechoslovakia. They are not 
developed in specimens recovered from the Spy Wood Grit which tend to have a 
limbus becoming obscure anteriorly and especially a dorsal beak located well forward 
of a rounded posterior margin. The Spy Wood sample is close to P. perfecta Cooper 
(1956 : 258) from the Pratt Ferry Formation of Alabama except in the strength of 
the limbus which is continuous in the American species. 

Despite these differences, the Shelve Paterula, provisionally at least, are most 
appropriately compared with P. bohemica and P. perfecta, the former to embrace 
specimens from the Mytton Flags to the Rorrington Beds inclusive, the latter shells 
occurring in the Spy Wood Grit and Aldress Shales. 



Order ACROTRETIDA Kuhn 1949 

Suborder ACROTRETIDINA Kuhn 1949 

Superfamily ACROTRETACEA Schuchert 1893 

Family ACROTRETIDAE Schuchert 1893 

APSOTRETA Palmer 1955 

Apsotreta sp. 

(PI. 6, figs. 14, 15) 

The complementary moulds of an acrotretid pedicle valve (BB 35565a, b), from 
the Mytton Flags exposed near Wood House (Grid Ref. SJ 338003), are the sole 



ORDOVICIAN BRACHIOPODA 43 

representatives of Apsotreta in the Shelve succession. The valve which had a 
subcircular commissure just under 3 mm in diameter was conical in profile and about 
two-fifths as deep as long with a planar apsacline pseudointerarea and an apical 
foramen ; the mould of the pseudointerarea was not sufficiently well preserved 
medially to determine whether an intertrough existed. The ornamentation con- 
sisted exclusively of fine growth lines. Internally the parallel-sided apical process, 
so diagnostic of the genus, extended forward for about two-fifths of the length of 
the anterior slope between 2 pairs of simply disposed vascula lateralia ; a pair of 
cardinal scars were impressed on the posterior surface. 

The valve cannot be profitably compared with those of other described Apsotreta, 
but it is noteworthy that its occurrence greatly extends the stratigraphic range of the 
genus which had previously been regarded as restricted to the Upper Cambrian 
(Rowell in Williams et al. : H276). 

CONOTRETA Walcott 1889 

Conotreta stapeleyensis sp. nov. 

(PI. 6, figs. 16-21) 

Diagnosis. Conotreta with a gently convex brachial valve averaging 85% as long 
as wide and a conical pedicle valve 37% as deep as long with a procline pseudo- 
interarea bearing a low deltoid arch ; apical process transversely oval, submedial 
vascula lateralia not deeply divided. 

Description. Medium-sized Conotreta with a very gently convex brachial valve 
averaging 85% as long as wide and 6% as deep as long (for 3 valves), and a conical 
pedicle valve averaging 37% as deep as long (for 3 valves) ; pseudointerarea well 
defined averaging 71% of the maximum width of 3 pedicle valves and procline or 
rarely nearly catacline in attitude, so that the beak with an apical foramen may be 
located forward of the posterior margin by as much as 20% of the valve length, 
pseudointerarea divided medially by low deltoid arch with a mean width of 30% 
of its length ; dorsal beak small, situated at the posterior margin defined by an 
anacline pseudointerarea ; external ornamentation consisting solely of fine growth 
lines. 

Ventral interior with a pair of subcircular cardinal scars impressed postero- 
laterally of a transversely oval apical process, mantle canal system pinnate with the 
submedial branches of the vascula lateralia not deeply divided. 

Dorsal interior with conspicuous propareas defined by a depressed median platform 
extending forward of the umbo for 18% of the length of the brachial valve and 
continuous with a long thin median septum of unknown profile ; cardinal and 
anterior muscle scars forming a quadripartite pattern about the median septum and 
divergent vascula lateralia and extending anteriorly of the umbo for 56% of the valve 
length. 

Type material. 

length width (mm) 
Holotype Internal mould of pedicle valve (BB 35566) 2-5 3-0 



44 SHELVE DISTRICT 

length width (mm) 

Paratypes External mould of brachial valve (BB 35567) 3-0 3-5 

Internal mould of brachial valve (BB 35568) 3-2 37 

Internal mould of brachial valve (BB 35569) 2-2 2*5 
Internal mould of brachial valve with 

adherent shell (BB 35570) 3-2 3-5 
External and incomplete internal mould of 

pedicle valve (BB 35571a, b) 3-5 37 

Horizon and locality. Stapeley Shales exposed on the road side north-east of 
bench mark 412-5, Leigh Hall (Grid Ref. SJ 334036). 

Discussion. A small number of acrotetid moulds recovered from the Stapeley 
Shales appear to constitute a new species of Conotreta. In the absence of precise 
data, no comparison can be drawn between the Stapeley Conotreta and other described 
species in respect of the shape of the shell and its principal morphological features. 
However, a combination of some of the basic characters of the genus immediately 
separate C. stapeley ensis from American (Cooper 1956 : 247-255), Baltic (Goryanski 
1969 : 62-65) an( i Scottish (Williams 1962 : 90) Conotreta. Such features include the 
development of a deltoid arch in place of an intertrough dividing a predominantly 
procline ventral pseudointerarea, and the late division of the submedial vascula 
later alia. 

The complementary moulds of a brachial valve which is likely to belong to 
Conotreta have also been found in the Hope Shales ; but, in the absence of an 
associated pedicle valve, its specific affinities cannot be determined. 



Superfamily DISCINACEA Gray 1840 

Family TREMATIDAE Schuchert 1893 

SCHIZOCRANIA Hall & Whitfield 1875 

Schizocrania salopiensis sp. nov. 

(PI. 6, figs. 22-26) 

Diagnosis. Schizocrania with subcircular brachial valve about one-third as 
deep as long with posteriorly placed umbo and dichotomizing radial capillae com- 
monly 8 to 10 per mm, 5 mm antero-medially of umbo ; large posterior adductors 
extending anteriorly for one-third the length of the valve. 

Description. Brachial valve subcircular and deep being 86% as long as wide 
and 33% as deep as long, transverse profile evenly convex, longitudinal profile 
asymmetrically convex with the posteriorly placed umbo overhanging a strong 
groove indenting the posterior part of the valve just within the smoothly rounded 
posterior margin ; exterior ornamented by sporadically occurring growth lines and 
capillae disposed radially from the umbo, curving to intersect the posterior margin, 
and branching by subequal dichotomy and commonly numbering 10 per mm, 5 mm 
antero-medially of the umbo ; dorsal interior with thickened rounded posterior 
margin extending postero-laterally for over four-fifths the width of the valve. 



ORDOVICIAN BRACHIOPODA 45 

Dorsal interior with faint median ridge ; posterior adductor scars large, suboval, 
slightly divergent and commonly with well-defined tracks, extending anteriorly of 
the posterior margin for 32% of the length of the valve ; anterior adductor scars 
small, suboval, normally obscure, situated about half-way along the valve and more 
medially placed than posterior pair ; a pair of short submedial fine ridges located 
forward of the anterior adductor scars may also represent muscle supports ; pedicle 
valve unknown. 

Type and figured material. 

length width (mm) 

Holotype External and internal moulds of brachial valve 

(BB 35358a, b) il-o 13-5 

Paratypes Exterior of brachial valve (BB 35359) 9-0 io-o 

External and internal moulds of brachial 
valve (BB 35360a, b) 1-9 2-3 

Exterior of brachial valve (BB 35361) 3-5 3-5 

External and internal moulds of deformed 
brachial valve (BB 35362a, b) 7-5 6-5 

External and internal moulds of brachial 
valve (BB 35404a, b) 4-2 5-0 

Horizons and localities. BB 35358, 35359 from Spy Wood Grit exposures 
1100 yds NNE of Rorrington (Grid Ref. SJ 303015) ; BB 35360 from Rorrington 
Beds exposed in Deadman's Dingle 100 ft upstream from its junction with Spy Wood 
Brook (Grid Ref. SO 280960) ; BB 35361, BB 35362 from Rorrington Beds exposed 
in tributary to Lower Wood Brook, 340 yds south-east of Desert (Grid Ref. 
SJ 308017) ; BB 35404a, b from Betton Beds exposed in a stream 330 yds WSW of 
Little Weston (Grid Ref. SO 290983). 

Discussion. Brachial valves of Schizomania occur rarely in the pre-gracilis 
rocks of the Shelve area, but are fairly common in the Rorrington Beds and Spy Wood 
Grit which have provided the two small samples used to derive the statistics given 
in Tables 27-29. Comparisons of these estimates of dorsal outline and depth and 
of the anterior extension of the posterior adductor scars show a significant difference 
(0-05 > p > 0-02) only in the relatively greater depth of the Spy Wood valves. 
This difference may reflect the growth of a consistently shallower valve during 
Rorrington time. Two other smaller samples from the Betton and Meadowtown 
Beds also differ from the Spy Wood specimens only in the depth of the brachial 
valve which averaged 19% and 20% of the length of 4 and 5 brachial valves respec- 
tively. However, many Schizocrania from the Betton, Meadowtown and Rorrington 
Beds show signs of collapse induced by compaction of the shaly sediments in which 
they occur ; and although there are no obvious signs of deformation in the specimens 
selected for measurement, the depressed condition of the valves which were mostly 
thin-shelled may yet prove to have been diagenetically induced. 

Apart from the information provided in Tables 27-29, the two samples are also 
alike in the rounded outline of the posterior margin and in details of the dichotomizing 
radial ornamentation. Thus between 4 and 5 mm anterior of the umbo, 7, 8, 9, 10, 



46 SHELVE DISTRICT 

II and 12 capillae per mm were counted respectively in o, i, o, 3, I and o specimens 
from the Spy Wood Grit and in 1, 3, 3, 5, 1 and 1 specimens from the Rorrington 
Flags. The ribbing of the Betton and Meadowtown specimens shows a similar range 
in frequency and, although 3 valves from the Hope Shales, with counts of 10, 12 
and 13 per mm, may be indicative of an older Schizomania with finer radial orna- 
mentation, the differences are not significant at present. 

The Shropshire Schizomania is closely related to 5. filosa (Hall) which is known 
principally from the younger Trentonian rocks of N. America (Cooper 1956 : 275). 
No details are available on the variation in the general morphology of the American 
species. But the Shropshire forms are distinguishable in possessing a rounded 
rather than truncated posterior margin and a dorsal umbo which is not marginal : 
and although no pedicle valves have yet been recovered, these differences in the 
brachial valve seem sufficiently important to warrant the erection of a new 
species. 

Table 27 

Statistics of the length (1) and maximum width (w) of n brachial valves from the Spy Wood Grit 
(A) and the Rorrington Beds (B) of Schizocrania salopiensis sp. nov. 





A 


B 


n 


6 


20 


1 mm (var 1) 
w mm (var w) 

r 
a (var a) 


6-42 (9-842) 

7-23 (15-01) 

0-997 

1-2349 (0-00259) 

Table 28 


4-29 (5-432) 

5-i (7-58) 

0-986 

1-1813 (0-0022) 



Statistics of the length (1) and depth (th) of n brachial valves from the Spy Wood Grit (A) and the 
Rorrington Beds (B) of Schizocrania salopiensis sp. nov. 





A 


B 


n 

1 mm (var 1) 

th mm (var th) 

r 
a (var a) 


7 

6-71 (8-822) 

2-24 (1-233) 

0-914 

0-3739 (0-0046) 


9 

4-i (3-74) 

077 (0-145) 

0-971 

0-1969 (0-00032) 



Table 29 

Statistics of the length (1) and length of the posterior adductor scars measured from the umbo 
(sc) of n brachial valves from the Spy Wood Grit and the Rorrington Beds of Schizocrania 

salopiensis sp. nov. 

A B 

n 4 10 

1 mm (var 1) 6-62 (14-23) 7-28 (12-984) 

s"c mm (var sc) 2-05(1-397) 2-07(1-193) 

r 0-997 0-952 

a (var a) 0-3133 (0-00033) 0-3031 (0-00107) 



ORDOVICIAN BRACHIOPODA 47 

Family DISCINIDAE Gray 1840 

Subfamily ORBICULOIDEINAE Schuchert & Le Vene 1929 

ORBICULOIDEA d'Orbigny 1847 

Orbiculoidea sp. 

(PL 7, fig. 1) 

An external mould with adherent shell of a brachial valve of Orbiculoidea (BB 35576) 
has been recovered from an unlocated exposure of Stapeley Shales. The valve 
which was 7 mm long and 1-2 mm deep was subcircular in outline with the beak 
situated about 2-5 mm forward of the slightly truncated posterior margin. The 
posterior sector was slightly concave, but the rest of the valve sloped evenly away 
from the beak to a flattened margin about 0-5 mm wide. The external surface was 
ornamented by fine fila ; no internal features are known. 

No useful purpose is served in comparing this brachial valve with those of 
described species, many of which are similarly proportioned. The specimen however 
constitutes an interesting record as one of the earliest Orbiculoidea so far identified. 

SCHIZOTRETA Kutorga 1848 

Schizotreta transversa sp. nov. 

(PI. 7, figs. 2, 3, 7) 

Diagnosis. Subcircular Schizotreta slightly wider than long with a ventral beak 
located one-third of the valve length forward of the posterior margin and a surface 
ornament consisting solely of growth lines. 

Description. Subcircular Schizotreta with a slightly truncated posterior margin, 
brachial valve very gently convex, pedicle valve subcorneal, almost nine-tenths as 
long as wide and just over one-fifth as deep as long, with an even slope away from 
the beak except for a slightly concave posterior margin and immediately in front of 
the slit -like pedicle opening ; dorsal beak submarginally located just over one-tenth 
forward of the posterior margin ; surface of both valves ornamented solely by fine 
growth fines ; interiors unknown. 

Type material. 

length width (mm) 

Holotype Exfoliated pedicle valve (BB 35577) 27 3-1 

Paratypes Exfoliated pedicle valve (BB 35578) 4-5 5-0 

Exfoliated brachial valve (BB 35579) 4-5 - 

Type horizon and localities. Meadowtown Beds : BB 35577, 35578 from 
exposures 80 yds from Meadowtown Quarry along the Rorrington road (Grid Ref . 
SJ 311012). BB 35579 from exposure in lane to Kinton, 200 yds west of Holy 
Trinity Church, Middleton (Grid Ref. SO 295993). 

Discussion. There seems to be little doubt that the discinid specimens collected 
from the Meadowtown Beds, and also a few less well-preserved moulds from the 
Rorrington Beds, should be assigned to Schizotreta because the beak of the brachial 



48 SHELVE DISTRICT 

valve is located submarginally. In other features, however, especially the trans- 
versely oval outline and the absence of fila, the new species is unlike penecontem- 
poraneous Schizotreta including the type species of the genus S. elliptica (Kutorga) 
from the Baltic area (Goryanski 1969 : 83), and Scottish and American stocks 
(Williams 1962 : 94-95 ; Cooper 1956 : 277-282). Only S. microthyris from the 
Oranda Formation of Virginia compares with S. transversa in outline although fila 
are finely but persistently developed on the American species and serve to distinguish 
the two. 

Schizotreta sp. 

(PL 7, %• 4) 

The external mould of a discinid pedicle valve (BB 35580), from the Mytton Flags 
exposed in the sides of the adit in the New Perkin's Level (Grid Ref. SJ 376022), has 
been provisionally identified as a Schizotreta rather than an Orbiculoidea mainly on 
the basis of its elongately oval outline and strong regular fila. The valve was 
4-5 mm long and about two-thirds as wide and one-fifth as deep as long. The beak 
was located about one-third the length of the valve anteriorly of the curved posterior 
margin with the pedicle opening represented by a narrow slit, about 0'5 mm long, 
breaching the slightly concave posterior surface. 

The valve is immediately distinguishable from 5. transversa in its elongately oval 
outline and strong fila which, together with the relatively anterior location of the 
beak and pedicle opening, render the specimen different from most other described 
species. However, in view of the fact that no Schizotreta has previously been reported 
from Lower Ordovician rocks (Rowell in Williams et al. 1965 : H285), more informa- 
tion, especially about the brachial valve, is required before taxonomic recognition 
is warranted. 

Suborder CRANIIDINA Waagen 1885 

Superfamily CRANIACEA Menke 1828 

Family CRANIIDAE Menke 1828 

PETROCRANIA Raymond 1911 

Petrocrania dubia sp. nov. 

(PI. 7, figs- 5, 6, 8) 

Diagnosis. Petrocrania known only by its subquadrate, conical brachial valve 
with the greatest depth corresponding more or less to mid-point ; elevated anterior 
adductor scars larger than posterior. 

Description. Exterior of brachial valve slightly irregular, subquadrate and 
suboval with flattened posterior margin, 82% as long as wide (mean for 4 valves) 
and 26% as deep as long (mean for 3 valves) with maximum depth at about mid- 
point ; surface with sporadic concentric growth lines, otherwise smooth ; shell 
with thickened slightly everted margin corresponding to limbus about 1 mm wide. 



ORDOVICIAN BRACHIOPODA 49 

Interior of brachial valve with strongly impressed suboval postero-lateral adductor 
scars smaller than suboval elevated anterior adductor scars situated submedially 
up to 60% forward of the posterior margin and ornamented by fine parallel wavy 
ridges ; mantle canal system pinnate with variably impressed sigmoidal vascula 
later alia. 

Type material. 

length width (mm) 
Holotype External and internal moulds of brachial valve 

(BB 35405a, b) n-5 14-0 

Horizon and locality. Whittery Shales : exposures in stream at north end of 
Spring Coppice 865 yds south-east of Hockleton Bridge (Grid Ref. SO 279997). 

Discussion. Although the sole craniid found in the Shelve area is represented by 
only 4 adult brachial valves from the Whittery Shales and a single immature brachial 
valve from the Aldress Shales, it is distinctive enough to be recognized as a new 
species. Its smooth slightly irregular external appearance suggests that it is best 
assigned to Petrocrania and the sigmoidal vascula lateralia impressed on the internal 
surface confirm this allocation. There are, however, two other internal features which 
are not typical of Petrocrania. They are the greater size of the anterior adductor 
scars compared with the posterior pair and the thickened margin. The new species, 
therefore, differs from other described Petrocrania in these characters. It is also 
distinctive in the mid-region location of the valve beak which is situated much nearer 
the posterior margin in American species of Ordovician age (Cooper 1956 : 287-291) 
and in P. inexpectata (Barrande 1848 : pi. no, fig. V) from the Caradocian Zahofany 
Formation of Czechoslovakia. 

Class ARTICULATA Huxley 1869 

Order ORTHIDA Schuchert & Cooper 1932 

Suborder ORTHIDINA Schuchert & Cooper 1932 

Superfamily ORTHACEA Woodward 1852 

Family HESPERONOMIIDAE Ulrich & Cooper 1936 

HESPERONOMIA Ulrich & Cooper 1936 

Hesperonomia sp. 

(PL 7, figs. 9, 12) 

The slightly deformed external and internal moulds (BB 35334a, b) of a brachial 
valve, collected from the Mytton Flags exposed between 50 and 55 yds down the 
lane from Woodside House (Grid Ref. SJ 337002), are representative of the genus 
Hesperonomia Ulrich & Cooper. The subquadrate gently concave valve was about 
7 mm long and had an anacline interarea about one-seventh as long as the valve, 
and an open notothyrium. The ornamentation consisted of fine costellae, arising by 
intercalation with a density of 6 per mm, 5 mm antero-medially of the umbo, and 
segregated into narrow sectors by a thickening of every eighth to tenth rib. In- 
ternally a blade-like cardinal process was supported on a low notothyrial platform 



50 SHELVE DISTRICT 

bounded by short divergent brachiophores and not prolonged anteriorly into a 
median ridge ; simple narrow sockets were obliquely developed between the hinge- 
line and the brachiophores. 

The combination of orthacean cardinalia, an unequally parvicostellate radial 
ornamentation and gently concave profile leave no doubt that the valve must be 
assigned to Hesperonomia as currently interpreted. Yet the ornamentation is so 
different from the normal costellate condition of the genus as to bring into question 
the use of the convexity or concavity of the brachial valve to distinguish between 
Hesperonomiella and Hesperonomia. Thus although the Mytton specimen is like 
Hesperonomia australis Ulrich & Cooper (1938 : 116) from the Odenville Limestone 
of Alabama, it more closely resembles the Arenigian Hesperonomiella carmalensis 
Bates (1968 : 142) from Anglesey in every respect except that the brachial valve of 
the Welsh species is very gently convex. Indeed, the Shropshire specimen may have 
been more closely related to H. carmalensis than to American Hesperonomia because 
the profile of the brachial valve was not as stable as its use for generic identification 
implies. 



Family ORTHIDAE Woodward 1852 

Subfamily ORTHINAE Woodward 1852 

ORTHIS Dalman 1827 

Orthis cf. callactis Dalman 1828 

(PL 7, figs. 10, 11, 13-16) 

Diagnosis. Transversely semi-oval, mucronate Orthis with a pedicle valve over 
one-third as deep as long and a planar brachial valve 57% as long as wide, orna- 
mented by about 17 costae with a wavelength of about o-8 mm, 5 mm antero- 
median^ of the dorsal umbo. 

Description. Plano-convex, transversely semi-oval Orthis with sharply acute 
cardinal angles and a pedicle valve, over one-third as deep as long, evenly convex in 
longitudinal profile but tending to become flattened postero-laterally ; mean length 
of 3 brachial valves 57% (range 52% to 60%) of the maximum width, slightly convex 
for the first 2 mm of growth with a shallow median sulcus but becoming plane and 
rectimarginate in later stages of growth ; ventral interarea short, about one-twelfth 
as long as the pedicle valve, strongly curved, apsacline ; dorsal interarea anacline, 
cardinal extremities acute at about 6o° ; ornamentation consisting of about 17 
angular costae, about 0-8 mm in wavelength 5 mm anterior of dorsal umbo, bearing 
fine lamellae and capillae ; postero-lateral sectors with about 5 short rugae disposed 
at an acute angle with the hinge-line. 

Ventral interior with teeth supported by short dental plates extending anteriorly 
for about one-seventh the length of the pedicle valve, pedicle callist small, muscle 
field with a rounded anterior boundary about one-quarter as long as the pedicle 
valve, adductor scar wide and not enclosed by the submedial diductor lobes. 



ORDOVICIAN BRACHIOPODA 51 

Dorsal interior with a simple cardinal process, short divergent brachiophores, and 
suboval adductor scars divided by a low median ridge and extending forward for 
about half the length of the brachial valve. 

Figured material. 

length width (mm) 
External moulds of conjoined valves (BB 35497) 5-2 8-0 

Incomplete external and internal moulds of pedicle valve 
(BB 35498a, b) 5-o 

Incomplete external and internal moulds of brachial valve 
(BB 35499) 2-0 

External moulds of conjoined valves (BB 35500) - 5-5 

Horizon and localities. Mytton Flags : BB 35498 from crag in the wood of 
Crowsnest Dingle, 500 yds north of Blakemoorflat (Grid Ref. SJ 374012) ; other 
specimens from exposures in the adit in Maddox's Coppice, 1200 yds ENE of St 
Luke's Church, Snailbeach (Grid Ref. SJ 382030). 

Discussion. The few moulds of Orthis s.s. recovered from the Mytton Flags are, 
when allowance is made for the small size of the shells which were rarely more than 
5-5 mm long, best identified as Orthis callactis Dalman, the type species for the genus. 
Illustrations of Russian representatives of the species from strata of B ni age 
(Schuchert and Cooper 1932 : PI. 2 ; Alichova 1953 : PL 1) show that adult shells 
were more subquadrate than semi-oval and the cardinal extremities less acute than 
those of the Mytton specimens. But growth lines indicate that during adult develop- 
ment there was an acceleration of forward growth relative to lateral increment and 
that the outline, profile and ornamentation of young shells rendered them in- 
distinguishable from the Shropshire specimens. 



Orthis sp. 

(PL 8, fig. 1) 

A few dorsal internal moulds from the Stapeley Volcanic Group, exposed 220 yds 
SSE of Whitsburn Cottage, Leigh (Grid Ref. SJ 332034), are best assigned to Orthis. 
The larger (BB 35425) which is 4 mm long and about 6-5 mm wide is the mould of a 
very slightly convex semicircular valve with squared cardinal angles which was 
ornamented by about 16 rounded costae and bore a linear cardinal process on a low 
notothyrial platform. The brachiophores, bounding the notothyrium and defining 
the narrow sockets, were short, widely divergent, and disposed at about the same 
angle as the almost hypercline interarea. 

The rounded costae, the slight convexity and the squared cardinal extremities 
suggest that the specimens are not assignable to 0. callactis Dalman. But it seems 
advisable to withhold systematic recognition until a sample of the pedicle valves 
has been recovered and more is known about the variability of the brachial 
valve. 



52 SHELVE DISTRICT 

LENORTHIS Andreeva 1955 

Lenorthis cf. proava (Salter) 

(PI. 8, figs. 2-9) 

1866 Orthis calligramma var. proava Salter : 336. 

1868 Orthis Carausii [Salter ms.] Davidson : 315. 

1869 Orthis Carausii [Salter ms.] Davidson : 229. 

1869 Orthis calligramma var. proava Salter ; Davidson : 241. 
1883 Orthis Carausii Salter ; Davidson : 182-184. 
1968 Lenorthis proava (Salter) Bates : 146. 

Diagnosis. Semi-oval Lenorthis with a brachial valve 70% as long as wide and 
17% as deep as long and a pedicle valve 31% as deep as long, ornamented by about 
19 costae with a wavelength of 075 mm, 5 mm antero-medially of the dorsal umbo, 
and fine capillae ; ventral muscle scar suboval extending anteriorly for 32% the 
length of the pedicle valve. 

Description. Ventribiconvex, transversely semi-oval Lenorthis with an evenly 
convex pedicle valve 31% as deep as long and a brachial valve 70% as long as wide 
and 17% as deep as long with a shallow sulcus extending antero-medially of the 
umbo for about 3 mm but becoming ill-defined thereafter ; ornamentation consisting 
of 15 to 20 narrowly rounded costae on 1, 1, o, 1, 5 and 1 brachial valves more than 
5 mm long, with a mean wavelength (and variance) of 0-75 mm (0-005), 5 mm anterior 
of the umbones of 9 brachial valves, bearing delicate capillae with those occurring 
medially in the interspaces tending to be stronger ; ventral interarea curved apsa- 
cline, about one-seventh as long as the pedicle valve ; dorsal interarea shorter 
anacline, cardinal extremities slightly acute or more rarely rectangular. 

Ventral interior with strong teeth supported by narrowly divergent dental plates 
extending anteriorly for 17% of the length of the pedicle valve, pedicle callist small, 
ventral muscle field subcordate, 96% as wide as long, with the submedial diductor 
lobes slightly longer than the broad median adductor scar and extending anteriorly 
for 32% of the length of the pedicle valve ; ventral mantle canal system rarely 
impressed but with divergent vascula media. 

Dorsal interior with ridge-like cardinal process, oblique sockets and short, rod-like 
brachiophores ; brachiophore bases extending anteriorly for about 14% of the length 
of the brachial valve and 49% as long as the lateral extension of the simple sockets ; 
dorsal adductor field quadripartite with the posterior pair of scars slightly larger, 
extending forward of the umbo for 56% of the length of the valve. 

Figured material. 

Internal and external moulds of pedicle valve (BB 35501a, b) 
Internal and external moulds of pedicle valve (BB 35502a, b) 
Internal and external moulds of brachial valve (BB 35503a, b) 
Internal and external moulds of brachial valve (BB 35504a, b) 

Horizon and localities. Mytton Flags : BB 35501, 35502 from exposures in 
roots of felled tree 70 yds north of entrance to Yewtree Level, the Hollies, Snailbeach 



length 


width (mm) 


6-5 


77 


57 


8-o 


5-3 


8-o 


7'5 


- 



ORDOVICIAN BRACHIOPODA 53 

(Grid Ref. SJ 380018) ; BB 35503, 35504 from crag in Perkin's Beach 170 yds west 
of Shepherd's Rock (Grid Ref. SO 372999). 

Discussion. The most remarkable feature of the two biconvex, costate orthids 
recovered from the Ordovician rocks of Shelve is that, although they belong to two 
different genera from widely separate horizons, they are indistinguishable in most of 
the attributes that are used to discriminate between species (Tables 30-38). Such 
dimensional estimates are numerical expressions of the way certain exoskeleton 
proportions were maintained during growth. They indicate the remarkable stability 
of the primitive strophic shell commonly referred to as 'typically orthid'. But 
there is no doubt that the two stocks can immediately be separated by the disposition 
of the ventral vascula media which is the chief diagnostic difference between Ortham- 
bonites and Lenorthis (Williams in Williams et al. 1965 : H311-313). 

Lenorthis proava was first described by Salter (1866 : 336) as a variety of Orthis 
calligramma Dalman from the Arenigian Carmel Grits of Anglesey. The type 
specimens have since been lost, but Bates (1968 : 147) has recently collected good 
samples from the same horizon. Comparisons between the Mytton and Anglesey 
shells show them to be the same in all morphological features except the costae 
which are significantly coarser in L. proava s.s. This difference alone is not con- 
sidered sufficiently important to warrant taxonomic recognition. It is, moreover, 
noteworthy that other penecontemporaneous forms from Europe and eastern North 
America may also prove to be morphologically comparable with L. proava. The 
accounts given by Rubel (1961 : 173) of ' Orthambonites' orbicularis (Pander) and of 
'Orthis' panderiana (Hall & Clarke) by Ulrich & Cooper (1938 : 102) suggest that 
these stocks are Lenorthis closely related to the British species and especially similar 
in the number of costae which appears to be the most reliable character for the differ- 
entiation of species. L. parvicrassicostatus (Cooper) from the Upper Ordovician 
of Scotland (Williams 1962 : 98) is also close to L. proava in the shape of its shell and 
the number of its costae, but the external surface is not capillate. 

ORTHAMBONITES Pander 1830 

Orthambonites exopunctata sp. nov. 

(PI. 8, figs. 10-17 '> see a l so Text-fig. 3, p. 24) 

Diagnosis. Subcircular Orthambonites with the brachial valve 74% as long as 
wide and 18% as deep as long and a pedicle valve 31% deep as long, ornamented by 
about 15 costae pierced by radial rows of exopuncta, and with a wavelength of o-8 mm, 
5 mm antero-medially of the dorsal umbo ; cordate ventral muscle scar extending 
anteriorly for 36% of the length of the pedicle valve, dorsal adductor scars impressed 
for 51% of the length of the brachial valve. 

Description. Ventribiconvex, subcircular Orthambonites with the pedicle valve 
31% as deep as long and evenly convex in transverse and longitudinal profiles ; 
brachial valve 74% as long as wide and 18% as deep as long with a shallow median 
sulcus almost half as wide as the length of the valve but tending to die out anteriorly ; 
ornamentation consisting of 14 to 16 subangular costae on 2, 4 and 3 brachial valves 



54 SHELVE DISTRICT 

more than 5 mm long, with a mean wavelength (and variance) of o-8i mm (0-005), 
5 mm anterior of the umbones of 7 brachial valves, crossed by fine lamellae and 
perforated by a row of exopuncta on either side of each costal crest in both valves ; 
ventral interarea curved, apsacline, about one-sixth as long as the pedicle valve ; 
dorsal interarea anacline, shorter, cardinal extremities invariably obtuse in shells 
longer than 3 mm. 

Ventral interior with blunt teeth rarely crenulated, supported by narrowly diver- 
gent dental plates extending anteriorly for 23% of the length of the pedicle valve, 
pedicle callist small, ventral muscle field cordate in adult valves 84% as wide as 
long with the submedial diductor lobes extending anteriorly for 36% of the length of 
the pedicle valve but not enclosing the shorter, medially divided, adductor scar ; 
ventral mantle canal system saccate with proximal parts of vascula media adjacent. 

Dorsal interior with ridge-like cardinal process, oblique sockets and short, rod-like 
brachiophores usually heavily encased in secondary shell, brachiophore bases 
extending anteriorly for 19% of the length of the brachial valve and 48% as long as 
the lateral extension of the simple sockets ; dorsal adductor field quadripartite 
extending forward of the umbo for 51% of the length of the valve ; dorsal mantle 
canal system not completely known but probably saccate. 

Type material. 

Holotype Exterior of brachial valve (BB 35371) 
Paratypes Exterior of brachial valve (BB 35372) 

Internal mould of brachial valve (BB 35373) 

Internal and external moulds of brachial valve 

(BB 35374a, b) 

Internal mould of pedicle valve (BB 35375) 

Pedicle valve and internal mould (BB 35376a, b) 

Horizon and localities. Whittery Shales : BB 35374 from exposures in the 
lower part of the west bank of the River Camlad 60 yds north-east of Marrington 
Farm (Grid Ref. SO 272970) ; BB 35375 from exposures in Whittery Quarry at the 
south end of Whittery Wood, near Chirbury (Grid Ref. SO 275981) ; all other 
specimens from outcrops in a stream at the north end of Spring Coppice 865 yds 
south-east of Hockleton Bridge, near Chirbury (Grid Ref. SO 279997). 

Discussion. Like Lenorthis proava the Whittery Orthambonites is characterized 
by a shell that did not deviate significantly during growth from the modal orthid 
shape and proportions, even to the development of obtuse cardinal extremities in 
shells only 2-5 mm long (Tables 30-38). Moreover, although the new species 
differs from most congeneric stocks, including the penecontemporaneous American 
0. bellus (Cooper 1956 : 296) and 0. friendsvillensis (Cooper 1956 : 303), in its sig- 
nificantly fewer ribs, in that character and all others that can be tested statistically 
it is indistinguishable from the Scottish Caradocian 0. playfairi (Reed emended 
Williams 1962 : 97). The Whittery shells, however, are unique among all known 
Orthambonites in bearing a row of exopuncta on either side of each costa of both 
valves. 



length 
8-5 


width (mm 

10-2 


7.4 
7-8 


9-0 
9-6 


4'5 
8-o 

8-5 


8-5 
9-0 



ORDOVICIAN BRACHIOPODA 55 

The exopuncta are shallow holes up to 0-2 mm deep, puncturing the external 
surface of the shell at acute angles directed posteriorly. Within 3 mm of the um- 
bones of each valve, the exopuncta tend to occupy the crests of costae. In larger 
shells they alternate on either side of the crest line at intervals of about 0-5 mm, but 
become more closely crowded and approximate to a double row in the marginal 
zones of shells more than 7 mm long. Judging from the inferred arrangement of the 
mantle canals relative to the costal embayments they probably accommodated 
setae which dehisced at regular intervals as the shell grew forward to seal them off 
from the mantle edge in the manner described by Rudwick (1965 : 604) for Acantho- 
thiris. If this interpretation is correct the arrangement and incidence of the 
exopuncta show that the growth of new setae not only balanced the loss of those 
isolated by shell deposition, but also compensated for the increasing expansion of the 
adult commissure as the decrease in forward growth tended to crowd the alternately 
arising setae into double rows which appeared more or less simultaneously. This 
pattern is different from that figured by Williams (in Williams et al. 1965 : H301) for 
Orthambonites cf. rotundiformis Cooper in which the morphology of the shell margin 
is compatible with the setal follicles having occupied narrow depressions in the 
intercostal eminences and not in the costal grooves. Such depressions are vestigial 
in 0. exopunctata ; but whether their presence is indicative of a more primitive 
anatomy as is suggested by their occurrence in Arenigian Orthambonites such as 0. 
calligramma Dalman remains to be demonstrated. 



Table 30 

Statistics of length (1) and maximum width (w) of n brachial valves of Orthambonites exopunctata 
sp. nov. (A) and Lenorthis cf. proava (Salter) (B) 

A B 

n 17 20 

1 mm (var 1) 4*97 (5-877) 4-8 (2-804) 

w mm (var w) 6-74 (8-36) 6-84 (3-479) 

r 0-984 0-867 

logel (var log e l) 1-497(0-2132) 1-512(0-1145) 

log e w (var log e w) 1-8226 (0-1695) 1-886 (0-072) 

r e 0-985 0-873 

a (var a) 0-892 (0-0016) 0-793 (0-00829) 



Table 31 

Statistics of length (1) and depth (th) of n brachial valves of Orthambonites exopunctata sp. nov. 

(A) and Lenorthis cf. proava (Salter) (B) 





A 




B 


n 


13 




8 


1 mm (var 1) 


6-55 (5-389) 




5-84 (1-569) 


th mm (var th) 


1-18 (0-271) 




0-99 (0-059) 


r 


0-863 




0-617 


a (var a) 


0-224 (o-oou 


■7) 


0-193 (0-003 



56 shelve district 

Table 32 

Statistics of length (1) and depth (th) of n pedicle valves of Orthambonites exopunctata sp. nov. (A) 

and Lenorthis cf . proava (Salter) (B) 





A 


B 


n 


24 


17 


1 mm (var 1) 


7-4 (5-892) 


5-51 (3-441) 


tn mm (var th) 


2-31 (0-674) 


1-72 (0-2106) 


r 


0-919 


0-578 


a (var a) 


0-338 (0-0008) 


0-247 (0-00272) 



Table 33 

Statistics of length of pedicle valve (1) and length of ventral muscle scar (sc) for n specimens of 
Orthambonites exopunctata sp. nov. (A) and Lenorthis cf. proava (Salter) (B) 

A B 

n 22 16 

1 mm (var 1) 7-79 (5-934) 57 1 (2-133) 

s"c mm (var sc) 2-77 (0-942) i-8i (0-343) 

r 0939 0-906 

a (var a) 0-399 (0-00094) 0-401 (0-00205) 



Table 34 

Statistics of length (1) and maximum width (w) of the ventral muscle scar of n specimens of 
Orthambonites exopunctata sp. nov. (A) and of Lenorthis cf. proava (Salter) (B) 





A 


B 


n 


18 


14 


1 mm (var 1) 


2-8i (0-789) 


1-78 (0-299) 


w mm (var w) 


2-36 (0-426) 


1-7 (0-174) 


r 


0-912 


0-895 


a (var a) 


0735 (0-00569) 


0-763 (0-00966) 



Table 35 

Statistics of length of pedicle valve (1) and length of dental plates (dl) for n specimens of Ortham- 
bonites exopunctata sp. nov. (A) and of Lenorthis cf. proava (Salter) (B) 

A B 

n 21 16 

1 mm (var 1) 7-23 (6-153) 5-75 (2-172) 

dl mm (var dl) i-68 (0-357) °'9% (°' I 37) 

r 0-923 0-895 

log e l (var log e l) 1-996 (0-098) 1-717 (0-0636) 

logedl (var log e dl) 0-46(0-1188) —0-66(0-1334) 

r e 0-926 0-903 

a (var a) 1-101(0-0091) 1-448(0-0275) 



ordovician brachiopoda 57 

Table 36 

Statistics of length of brachial valve (1) and length of brachiophores at their bases (c) for n 
specimens of Orthambonites exopunctata sp. nov. (A) and of Lenorthis cf. proava (Salter) (B) 





A 


B 


n 


8 


13 


1 mm (var 1) 


6-i6 (4-389) 


5-6 (1-833) 


c mm (var c) 


1-19 (0-167) 


o-8 (0-032) 


r 


0-971 


0-729 


a (var a) 


0-195 (0-00036) 


0-132 (0-00074) 



Table 37 

Statistics of length of brachiophores at their bases (1) and distances between the lateral edges of 
sockets (w) for n brachial valves of Orthambonites exopunctata sp. nov. (A) and Lenorthis cf. 

proava (Salter) (B) 

A B 

n 6 12 

1 mm (var 1) 1-03(0-111) 0-81(0-034) 

w mm (var w) 2-15 (0-347) I- ^7 (0-09) 

r 0-944 0-727 

a (vara) I'll 1 (0-0848) 1-637 (0-1262) 

Table 38 

Statistics of length of brachial valve (1) and length of adductor scar from umbo (sc) of n specimens 
of Orthambonites exopunctata sp. nov. (A) and of Lenorthis cf. proava (Salter) (B) 





A 


B 


n 


8 


7 


1 mm (var 1) 


6-i6 (4-389) 


5-46 (2-913) 


sc mm (var sc) 


3-13 (0-899) 


3-06 (0-767) 


r 


0-967 


0-948 


a (var a) 


0-453 (0-00221) 


0-513 (0-0054) 



Subfamily PRODUCTORTHINAE Schuchert & Cooper 1931 

NICOLELLA Reed 1917 

Nicolella cf. actoniae (J. de C. Sowerby) 

(PI. 9, figs. 1-6) 

1839 Orthis actoniae J. de C. Sowerby in Murchison : 639. 
1963 Nicolella actoniae (Sowerby) Williams : 352. 

Diagnosis. Plano-convex, mucronate Nicolella with a pedicle valve 32% as 
deep as long and a brachial valve 64% as long as wide, ornamented by 12 to 15 
angular costae with a wavelength of about i-i mm, 5 mm antero-medially of the 
dorsal umbo ; costellae rare with 3a arising first, cordate ventral muscle scar extend- 
ing anteriorly for 34% of the length of the pedicle valve, dorsal adductor scars 
impressed for 48% of the length of the brachial valve. 



58 SHELVE DISTRICT 

Description. Piano- to slightly concavo-convex, subquadrate Nicolella with 
cardinal angles almost invariably acute in all growth stages, pedicle valve evenly 
convex 32% as deep as long and a brachial valve 64% as long as wide ; ornamenta- 
tion consisting of 12 to 15 angular costae in 2, 1, 2 and 2 brachial valves more than 
8*5 mm long with a mean wavelength (with variance) of i-og mm (0-008) at 5 mm 
antero-medially of the umbones of 10 brachial valves, rarely branching in shells 
longer than 7-5 mm with 3a usually arising first, concentric lamellae strong and 
continuous but not occurring regularly ; ventral interarea short, curved orthocline ; 
dorsal interarea very short, curved anacline. 

Ventral interior with small blunt teeth supported by widely divergent dental 
plates extending anteriorly for 16% of the length of the pedicle valve, pedicle callist 
small, ventral muscle field cordate, slightly wider than long in adult valves with the 
submedial diductor lobes extending anteriorly for 34% °f the length of the pedicle 
valve but not enclosing the shorter, undifferentiated median adductor scar. 

Dorsal interior with ridge-like erect cardinal process, chilidial plates present, 
sockets slot-like defined by brachiophores extending laterally parallel to the hinge- 
line for a distance equivalent to about one-third of the length of the brachial valve ; 
dorsal adductor scars generally obscure but quadripartite with the larger anterior 
pair extending forward of the dorsal umbo for 48% of the length of 3 valves (range 
46% to 50%). 

Figured material. 

length width (mm) 

External and internal moulds of brachial valve (BB 35327a, b) 3-0 5-5 

External and internal moulds of brachial valve (BB 35328a, b) io-o 15-0 

External mould of brachial valve (BB 35329) 13-0 18-0 
Incomplete internal mould of pedicle valve (BB 35330) - - 

Internal mould of pedicle valve (BB 35331) 12-0 15-0 

Horizon and localities. Whittery Shales : BB 35327 from Whittery Quarry 
at south end of Whittery Wood near Chirbury (Grid Ref. SO 275981) ; BB 35328 
from outcrops in west bank of the River Camlad 60 yds north-east of Marrington 
Farm (Grid Ref. SO 272970) ; BB 35329, 35331 from outcrops half way down the 
path going south from Marrington Farm to the valley bottom (Grid Ref. SO 272967) ; 
BB 35330 from outcrops in stream at north end of Spring Coppice 865 yds south- 
east of Hockleton Bridge (Grid Ref. SO 279997). 

Discussion. With the identification of the Whittery Nicolella as a form which is 
indistinguishable from the Actonian N. actoniae (J. de C. Sowerby), the remarkable 
morphological stability of that species becomes apparent, for it extends throughout 
the Upper Caradocian and Lower Ashgillian without significant change in any of its 
definable features (Williams 1963 : 352-356, Wright 1964 : 165-167). Tables 39-42 
give the statistics for the main external and internal features of the Whittery speci- 
mens, and their comparison with similar data for N. actoniae from the Acton Scott 
Beds and the Portrane Limestone, and for N. actoniae obesa Williams from the 
Gelli-grin Calcareous Ashes, shows no significant difference in any of the listed 
characters. It is noteworthy that even in respect of the relative depth of the pedicle 



ordovician brachiopoda 59 

Table 39 

Statistics of length (1) and maximum width (w) of 9 brachial valves of Nicolella cf. actoniae 

(Sowerby) 

1 mm (var 1) 7-93 (12-915) 

w mm (var w) 12-37 (26-845) 

r 0-978 

a (var a) I '44 I 7 (0-0128) 

Table 40 

Statistics of length (1) and maximum depth (th) of 15 pedicle valves of Nicolella cf. actoniae 

(Sowerby) 

1mm (var 1) 11-53 (12-595) 

th mm (var th) 3-66 (1-41) 

r 0-854 

a (var a) 0-3346 (0-00234) 

Table 41 

Statistics of length of pedicle valve (1) and length of ventral muscle scar (sc) for 12 specimens of 

Nicolella cf. actoniae (Sowerby) 

1 mm (var 1) 12-79 (3-019) 

s"c mm (var sc) 4-29 (0-481) 

r 0-802 

a (var a) 0-3991 (0-00568) 

Table 42 

Statistics of length of pedicle valve (1) and anterior extension of dental plates (dl) of 7 specimens 

of Nicolella cf . actoniae (Sowerby) 

1 mm (var 1) H'73 (6-778) 

dl mm (var dl) 1-91 (0-332) 

r 0964 

a (var a) 0-2212 (0-0007) 

valve the Whittery sample is intermediate between N. actoniae and the significantly 
deeper Bala subspecies and, when allowance is made for the variability of the 
samples, comparable with both in the inherent profile and rate of deepening of the 
pedicle valve. 

The development of the ribbing does, however, indicate the much closer affinity 
with N. actoniae than with the Welsh subspecies. In the number and wavelength 
of costae the Whittery sample is like both N. actoniae and its subspecies. However, 
in all 8 brachial valves between 8-5 and 14-2 mm long, no costellae arise within 
7-5 mm of the umbo and only 3 out of 8 valves show any branching, which consists 
solely of the appearance of 3a in two specimens and 4a arising before 3a in the third. 
The Whittery sample is therefore more like N. actoniae in the rare development of 
costellae within the size range stipulated above, and also in being less thick-shelled 
than the Bala subspecies at such stages in growth (Williams 1963 : 356). 



60 SHELVE DISTRICT 

Only a few other remarks can be made about the Whittery shells. As in other 
Nicolella, muscle impressions are generally poorly defined. The ventral muscle 
field was wider than long in young shells, but during growth there was an accelera- 
tion in the anterior extension of the submedial diductor lobes so that the muscle 
field of adult pedicle valves tended to be longer than wide ; and the mean percentage 
width relative to length from the umbo (with variance) of 6 muscle impressions 
between 3-0 and 5-5 mm long was 109-4% (425-84). 



Subfamily WHITTARDIINAE nov. 

Diagnosis. Piano- to concavo-convex orthids with a catacline to procline ventral 
interarea, disharmonic costellate ornamentation and open delthyrium and noto- 
thyrium ; ventral muscle field strongly bilobed with short parallel dental plates 
and three or four pairs of short ridges radiating from the muscle field boundary ; 
cardinalia consisting of a simple cardinal process and widely divergent brachiophores 
ankylosed to the hinge line, adductor scars subtriangular, divided into inner and 
outer pairs by vascula myaria ; ventral mantle canal system saccate, dorsal lem- 
niscate to pinnate. 

Discussion. The new genus Whittardia represents an unexpected find within 
such well-known Ordovician successions as those of Shropshire, because although it 
is undoubtedly orthacean in its general organization and more particularly orthid 
in the bilobed nature of the ventral muscle field and the simplicity of the cardinalia, 
it is sufficiently unusual in a number of features to have aroused the attention of 
palaeontologists had it been discovered earlier. Among such features, the dis- 
harmonic ornamentation of the valves in the earlier stages of growth is the most 
bizarre. Judged on the umbonal region of adult moulds, dorsal valves, up to 
about 5 mm long, were convex and sulcate and ornamented by three symmetrically 
disposed pairs of strong costae radiating from the umbo. The only other ornamenta- 
tion developed within 2 mm of the umbo, in all valves examined, consisted of fine 
closely-spaced lamellae. Thereafter at variable distances from the umbo inter- 
calated costellae appeared, normally with a few thickened differentially. But with 
the relative loss in strength of the primary costellae in valves longer than 5 mm, the 
pattern became more evenly multicostellate with the costae arising by dichotomy or 
intercalation and interrupted by fine concentric lamellae which were especially 
conspicuous in the postero-lateral sectors free of ribs. The umbonal surface of the 
pedicle valve, on the other hand, appears at first sight to have been quite different, 
although details are usually obscured by the presence of a large mesothyridid fora- 
men and the irregularity of the surface. The first-formed part of the preserved 
external mould, for about 2 mm anterior of the foraminal edge, appears to be flat 
and smooth and although valves up to about 5 mm long were sufficiently carinate to 
accommodate the dorsal sulcus they were evenly and closely costellate and without 
sign of accentuated interspaces corresponding to the strong costae of the brachial 
valve. In the ventral interior, however, there were three or four pairs of low 
narrow ridges radiating from the boundaries of the ventral muscle field and dying 



ORDOVICIAN BRACHIOPODA 



61 



out peripherally, which could well have been complementary to the accentuated 
costae of the brachial valve. This arrangement suggests that the valves were 
imperfectly matched along the commissure, and such a loose fit is compatible with 
the weak articulation indicated by the simplicity of the teeth, the shallowness of the 
sockets, and the powerful development of the adductor musculature as inferred from 
the wide ventral impression and the deep posterior insertion of the dorsal attachment 
areas. 

The other unusual features of the new genus are the catacline to procline ventral 
interarea and the profile of the brachial valve. The disposition of the ventral 
interarea was not unique among orthaceans but it was associated with a large 
mesothyridid foramen which must have accommodated a thick pedicle that was so 
short as to bring the pedicle valve into contact with the substratum ; the irregular 
contours of the valve are not post mortem features but expressions of growth in a 
physically restrictive environment. The longitudinal profile of the brachial valve 
was initially convex, but beyond the 2 mm growth stage the valve became gently 
concave. This attitude, however, was accentuated or reversed along a few narrow 
arcs resembling concentric rugae or, if incomplete, incipient geniculation. Such 
bands corresponded to only vague concentric indentations in the ventral interior, 
again indicative of the degree to which the valves grew independently of each other 
at the edges. 

No other described orthacean compares closely with Whittardia : indeed the 
principal difficulty is deciding the suprageneric group to which it is best assigned. 
There is a superficial resemblance to the dinorthid Plesiomys (Retrosistria) in the 
ventral muscle impression but this is due to the catacline attitude of the ventral 
interarea in both stocks. Yet the simplicity of the cardinalia, and the bilobed nature 
of the ventral muscle scar and the mantle canal systems, suggest that the ancestor 
of Whittardia was more likely to have been an orthid and, provisionally at least, a 
monotypic subfamily of the Orthidae has been created to accommodate the new genus. 




tooth 

cardinal process 

socket 

brachiophore 
dental plate 

muscle scar 
mantle cana 




Fig. 5. Diagrammatic views of (A) the ventral and (B) the dorsal interiors of 

Whittardia. 



62 SHELVE DISTRICT 

Genus WHITTARDIA nov. 

Name. In honour of the late Professor W. F. Whittard. 

Diagnosis. Subquadrate to semi-elliptical, sulcate shell with an irregularly 
convex, carinate pedicle valve and a brachial valve initially convex but becoming 
concave in adult stages of growth ; radial ornamentation of both valves costellate, 
interrupted by fine lamellae, but with the finely lamellose umbonal region of the 
brachial valve bearing only 6 symmetrically disposed accentuated costae becoming 
subdued anteriorly ; ventral interarea gently curved catacline to procline, with 
wide open delthyrium and mesothyridid foramen, dorsal interarea very short, 
anacline, notothyrium open ; shell probably impunctate. 

Ventral interior with widely divergent tooth ridges and flat, oblique teeth supported 
by short receding subparallel, widely spaced dental plates, postero-laterally enclosing 
a bilobed muscle field with a relatively short undifferentiated adductor scar separat- 
ing divergent submedial diductor lobes which encroach onto a pair of divergent 
vascula media ; ventral mantle canal system probably saccate with vascula media 
branching early ; 3 or 4 pairs of narrow, low ridges radiate from the boundary of the 
muscle field and die away peripherally. 

Dorsal interior with simple cardinal process elevated above a shallow notothyrial 
platform and wide short median ridge ; brachiophores short, widely divergent, 
ridge-like, scarcely elevated above hinge-line and postero-laterally defining a pair 
of elliptical sockets ; adductor scars subtriangular with posterior apices deeply 
inserted in the notothyrial platform and divided into outer and inner sets by obliquely 
disposed vascula my aria ; dorsal mantle canal pattern lemniscate or pinnate. 

Type species. Whittardia paradoxica sp. nov. from the Whittery Shales, 
Shropshire. 

Whittardia paradoxica sp. nov. 
(PL 9, figs. 7-15 ; PI. 10, figs. 1, 2, 4 ; Text-fig. 5) 

Diagnosis. Concavo-convex, subquadrate sulcate Whittardia with a pedicle 
valve 34% as deep as long and a brachial valve 79% as long as wide, ornamented 
by multicostellae with counts of 3 or 4 ribs per mm, 5 mm antero-medially of the 
dorsal umbo ; ventral and dorsal muscle scars extending anteriorly for 24% and 
46% of their respective valve lengths. 

Description. Variably subquadrate, sulcate Whittardia with a convex medially 
carinate pedicle valve with a mean length relative to width of 66% (range 56% to 
83%) and a mean depth relative to length of 34% (range 26% to 41%) for 4 valves, 
and an irregularly concave brachial valve averaging 79% as long as wide (range 71% 
to 84%) for 4 valves ; radial ornamentation becoming multicostellate by inter- 
calation and dichotomy in adult shells with 3 and 4 ribs per mm, 5 mm antero- 
medially of the umbones of 2 brachial valves in each case. 

Ventral interior with bilobed ventral muscle scar becoming longer than wide in 
adult shells and extending forwards for an average of 24% of the length of 4 pedicle 
valves (range 20% to 33%). 



ORDOVICIAN BRACHIOPODA 63 

Dorsal interior with cardinalia consisting of a simple cardinal process and widely 
divergent brachiophores extending forwards for an average of 17% (range 14% to 
19%) of the length of 4 pedicle valves and supported by a shallow notothyrial plat- 
form ; dorsal adductor scars subtriangular, deeply impressed posteriorly on either 
side of a short low median ridge, and extending forwards for 46% (range 42% to 
51%) of the length of 4 brachial valves. 

Type material. 

length width (mm) 

Holotype Internal and external moulds of brachial valve 

(BB 35381a, b) 6-5 7-5 

Paratypes Internal and external moulds of pedicle valve 

(BB 35382a, b) n-5 14-0 

Internal and external moulds of pedicle valve 

(BB 35383a, b) 7-5 

Internal and external moulds of brachial valve 

(BB 35384a, b) 8-o 

Internal mould of brachial valve (BB 35385) 7-2 - 

Internal mould of pedicle valve (BB 35386) 7-5 - 

External and internal moulds of brachial valve 

(BB 35402a, b) 7-0 

External and internal moulds of pedicle valve 

(BB 35403a, b) 7-0 

Horizon and localities. Whittery Shales : BB 35384 from outcrops in stream 
at north end of Spring Coppice, 865 yds south-east of Hockleton Bridge (Grid Ref. 
SO 279997) ; all other specimens from Whittery Quarry at the south end of Whittery 
Wood, near Chirbury (Grid Ref. SO 275981). 

Discussion. The sample of Whittardia paradoxica available for assessing the 
morphological variation within the species was too small to estimate proportional 
growth, but statistics incorporated in the description were calculated for valves 
between 3 and 12-4 mm in length and give some idea of the variability. 

Family DOLERORTHIDAE Opik 1934 

Subfamily DOLERORTHINAE Opik 1934 

DOLERORTHIS Schuchert & Cooper 1931 

Dolerorthis cf. tenuicostata Williams 

(PL 10, figs. 3, 5, 7, 10, 13) 

I 955 Dolerorthis tenuicostata Williams in Whittington & Williams : 406. 

The incomplete external and internal moulds of a brachial and pedicle valve 
(BB 35459a, b ; BB 35460a, b) of Dolerorthis have been collected from the Whittery 
Shales exposed in Whittery Quarry in the south end of Whittery Wood, near 
Chirbury (Grid Ref. SO 275981). The shell was ventri-biconvex with a pedicle 
valve about one-fifth as deep as long, sloping evenly away from the beak, and a very 



64 SHELVE DISTRICT 

gently convex, transversely semi-oval brachial valve just over two-thirds as long as 
broad and lacking a sulcus. The ventral interarea was strongly apsacline and, at 
one-eighth the length of the pedicle valve, longer than the anacline dorsal interarea ; 
both delthyrium and notothyrium were open. The ornamentation consisted of 
microscopic concentric lamellae superimposed on fine costellae numbering 4 and 5 
per mm at 5 mm antero-medially of the umbones, with costellae, mainly branching 
internally, arising early so that about 60 occur at the 5 mm growth stage and 75 at 
the margin of the brachial valve 8-5 mm long. 

A pair of dental plates extending forward for about one-seventh the length of the 
pedicle valve defined the umbonal cavity which was occupied by a subtriangular 
muscle field with a relatively broad median adductor track ; only proximal traces of 
the vascula media, separated by a short median ridge, are preserved. 

The cardinalia consisted of a narrow ridge-like cardinal process and sharp, 
divergent, unsupported brachiophores extending forward for one-fifth of the length 
of the brachial valve ; only faint impressions of a pair of suboval adductor scars are 
preserved on either side of a low median ridge becoming indistinct anteriorly. 

The fineness of the ribbing as well as the internal features indicate that the 
Whittery Dolerorthis resembles D. tenuicostata (Williams 1955 : 406) from the Derfel 
Limestone of N. Wales. There are, however, some differences especially in the 
lesser convexity of the brachial valve of the Whittery species which also lacks a 
sulcus. These differences may ultimately prove to be taxonomically important. 



Subfamily GLYPTORTHINAE Schuchert & Cooper 1931 

GLYPTORTHIS Foerste 1914 

Glyptorthis viriosa sp. nov. 

(PI. 10, figs. 6, 8, 9, 11, 12, 14, 15 ; PI. 11, figs. 1, 2, 4) 

Diagnosis. Small Glyptorthis with a shallowly sulcate brachial valve about four- 
fifths as long as wide and a catacline pedicle valve 37% as deep as long, ornamented 
by fascicostellae numbering 2 or 3 per mm, 5 mm antero-medially of the dorsal 
umbo, and concentric lamellae with a modal density of 3 per mm between 5 and 6 mm 
antero-medially of the dorsal umbo. 

Description. Subcircular, ventribiconvex Glyptorthis with roundly obtuse 
cardinal angles ; pedicle valve sharply carinate and 37% as deep as long, brachial 
valve 80% as long as wide and 23% as deep as long with a persistent rounded median 
sulcus, up to half as wide as the valve length, flanked by evenly convex lateral areas ; 
ventral interarea catacline to slightly apsacline, about one-third as long as the valve, 
with narrow delthyrium with subparallel boundaries ; dorsal interarea anacline 
with an open notothyrium ; radial ornamentation fascicostellate with ribs branching 
almost exclusively internally in the first four sectors and numbering 2 and 3 per mm, 
5 mm antero-medially of the umbones of 3 brachial valves in each case ; concentric 
ornamentation consisting of strongly developed lamellae with counts of 3 and 4 in 
3 and 1 valves between 5 and 6 mm antero-medially of the umbones. 



ORDOVICIAN BRACHIOPODA 65 

Small trigonal teeth supported by narrowly divergent dental plates extending 
forward for 20% of the length of the pedicle valve ; ventral muscle scar normally 
elongately pentagonal, with relatively wide median adductor scars flanked by some- 
what shorter diductor scars, wider than long in young valves but becoming much 
more elongate in adult shells with a mean width of 97% of the length and extending 
anteriorly for 38% of the length of the valve. 

Cardinal process simple and blade-like, brachiophores divergent and pointed with 
the secondary shell, forming the boundaries of the denticulate sockets and noto- 
thyrium, elevated above the brachial valve floor to simulate fulcral plates ; convergent 
brachiophore bases extending forward for 19% of the valve length and 48% as long 
as wide ; adductor muscle field quadripartite with a pair of suboval scars impressed 
on either side of the median ridge and extending anteriorly of the umbo for 52% of 
the length of the valve. 

Type material. 

length width (mm) 

Holotype External and internal moulds of brachial valve 

(BB 35505a, b) 47 6-0 

Paratypes External and internal moulds of pedicle valve 

(BB 355o6a, b) 4-5 5-5 

Internal mould of pedicle valve (BB 35507) 5-0 6-5 

Internal and incomplete external moulds of 
brachial valve (BB 35508a, b) 5-5 7-5 

Internal mould of pedicle valve (BB 35509) 6-5 7-0 

Type horizon and locality. All specimens from exposures of Spy Wood Grit, 
1100 yds NNE of Rorrington (Grid Ref. SJ 303015). 

Discussion. Remains of Glyptorthis have been recovered from the Meadowtown 
Beds, Whittery Shales and Spy Wood Grit, but are common only in the last-named 
formation which has furnished the sample used in establishing a new species most 
closely related to G. nantensis McGregor (1961 : 187) from the Upper Llandeilo of 
the Berwyn Hills. However, even in the absence of precise information about the 
variability of the Welsh species, its pedicle valve is obviously not only much deeper 
but is further characterized by an interarea which is 'slightly apsacline, almost 
orthocline', and a significantly larger muscle scar. 

Other penecontemporaneous Ordovician species from Ireland, Scotland and N. 
America also differ greatly. According to data kindly provided by Mrs H. Carlisle, 
in the brachial valve of G. crispa M'Coy from the Tramore Limestone of Ireland the 
cardinalia were relatively wider while the ventral muscle scar extended anteriorly 
more slowly during growth than in the Shropshire species. G. balclatchiensis 
(Davidson) from the Ardmillan Series of Scotland (Williams 1962 : 109), and G. 
assimilis Cooper (1956 : 361) from the Lower Ridley of Tennessee, both differ in 
being significantly wider. The former species also differs in the replacement of the 
dorsal sulcus by a narrow rounded fold during adult stages of growth, and the latter 
in being more coarsely imbricate. 



66 SHELVE DISTRICT 

Specimens of Glyptorthis recovered from the Whittery Shales and Meadowtown 
Beds may, provisionally at least, be referred to the new species despite minor 
differences in morphology. 

The few moulds known from the younger formation include only one external 
mould of a brachial valve large enough to provide data about ornamentation. 
Counts of 2 per mm at 5 mm antero-medially of the umbo were obtained for both 
radial and concentric ornamentation and are smaller than those for G. viriosa s.s. In 
all other features, however, the shells must have been indistinguishable from the 
Spy Wood forms and more information is necessary to establish the significance of 
the difference. 

The more frequently occurring Glyptorthis from the Meadowtown Beds, on the 
other hand, is identical with the new species except for the size of the cardinalia 
which extend anteriorly for only 10% of the length of the two known brachial valves. 
This size relationship is subnormal for brachial valves of comparable length in the Spy 
Wood sample, although more specimens are needed to assess the significance of the 
difference. Two specimens are figured (PI. 10, fig. 15 ; PL 11, figs. 1, 2, 4) : a 
brachial valve (BB 35520a, b) from loose flags by the side of Minicop Farm (Grid 
Ref. SJ 314018), and a pedicle valve (BB 35521a, b) from Quinton's Quarry in the 
field 200 yds north-east of Meadowtown Chapel (Grid Ref. SJ 312013). 



Table 43 

Statistics of length (1) and maximum width (w) of 17 brachial valves of Glyptorthis viriosa sp. nov. 

1 mm (var 1) 4-23 (3-037) 

w mm (var w) 5-26 (3-424) 

r 0-978 

logel (var log e l) 1-3637 (° -I 567) 

logeW (var log e w) 1-6029 (0-1165) 

r e 0-981 

a (var a) 0-8621 (0-00191) 

Table 44 

Statistics of length (1) and maximum depth (th) of 19 brachial valves of Glyptorthis viriosa sp. nov. 

1 mm (var 1) 4-28 (3-05) 

th mm (var th) 0-97 (0-162) 

r o-866 

a (var a) 0-2306 (0-00078) 

Table 45 

Statistics of length (1) and maximum depth (th) of 23 pedicle valves of Glyptorthis viriosa sp. nov. 

1 mm (var 1) 3-46 (4-857) 

th mm (var th) 1-27 (0-594) 

r 0-945 

a (var a) 0-3497 (0-00061) 



ordovician brachiopoda 67 

Table 46 

Statistics of length (1) and length of dental plates (dl) in 20 pedicle valves of Glyptorthis viriosa 

sp. nov. 

1 mm (var 1) 3-94 (4-674) 

dl mm (var dl) o-8i (0-246) 

r 0-952 

a (var a) 0-2293 (0-00027) 

Table 47 

Statistics of length (1) and length of ventral muscle scar (sc) in 22 pedicle valves of Glyptorthis 

viriosa sp. nov. 

1 mm (var 1) 3-94 (4-848) 

sc mm (var sc) 1*49 (1-124) 

r 0-989 

log e l ( var logel) 1-237(0-2711) 

log e sc (var log e sc) 0-1948 (0-4092) 

r e 0-994 

a (var a) 1-2285 (0-00095) 

Table 48 

Statistics of length (1) and width (w) of the ventral muscle scar in 23 pedicle valves of Glyptorthis 

viriosa sp. nov. 

1 mm (var 1) 1-46 (1-088) 

w mm (var w) 1-42 (0-282) 

r 0-932 

logel ( var logel) 0-1772(0-4098) 

logeW (var log e w) 0-2866 (0-1305) 

re 0-957 

a (var a) 0-5642 (0-00126) 

Table 49 

Statistics of length (1) and length of brachiophore bases (lc) in 19 brachial valves of Glyptorthis 

viriosa sp. nov. 

1 mm (var 1) 4-62 (2-227) 

lc mm (var lc) o-86 (0-084) 

r 0-865 

a (var a) 0-1938 (0-00055) 

Table 50 

Statistics of length (1) and maximum lateral extension (w) of the brachiophore bases in 18 brachial 

valves of Glyptorthis viriosa sp. nov. 

1 mm (var 1) o-86 (o-o88) 

w mm (var w) 1-78 (0-274) 

r 0-842 

a (var a) 1-762 (0-05665) 



68 SHELVE DISTRICT 

Family ALIMBELLIDAE Andreeva i960 emended A. W. 

Diagnosis. Biconvex, smooth to capillate, uniplicate orthaceans with well- 
developed curved interareas and open delthyrium and notothyrium ; ventral muscle 
field impressed on pseudospondylium, dental plates absent at least in adult shells ; 
blade-like cardinal process and strong brachiophores supported by variably devel- 
oped notothyrial platform ; ventral mantle canal pattern saccate, dorsal digitate. 

Discussion. This family was erected by Andreeva (i960 : 292) to include two 
Tremadocian genera from the Urals, Alimbella Andreeva i960 and Medesia An- 
dreeva i960 which were considered by her to be aberrant porambonitaceans, a 
conclusion accepted by Biernat (in Williams et al. 1965 : H530), in her review of the 
Syntrophiidina for the Treatise on Invertebrate Paleontology. In her discussion of 
the affinities of the Russian genera, Andreeva conceded that the morphology of the 
brachial valves is typically 'orthoid, similar to Platystrophia (i960 : 292) but believed 
the organization of the pedicle valve to be so decisively porambonitacean as to 
preclude any connection with the orthaceans. It is difficult to understand how 
Andreeva came to this conclusion unless she had attached an overriding importance 
to the uniplicate condition of the shell which is certainly much more characteristic 
of the porambonitaceans than the orthaceans. Yet, if the ventral sulcus is ignored, 
the pedicle valves of both genera are immediately seen to be unexceptionally 
orthacean in the strong development of the interareas, the arrangement of the ventral 
muscle fields, and above all in the saccate mantle canal systems, with the proximal 
parts of the vascula media disposed medially in a manner which is unknown among 
the digitate porambonitaceans. The pseudospondylium is admittedly a rare feature 
among orthaceans but is known to have developed independently in several articulate 
groups and is, in any event, as well defined in the orthid Glossorthis as it is among the 
alimbellids. 

All other features displayed by the Russian genera, such as the cardinal process 
and lack of brachiophore supports, are much more, but not exclusively, characteristic 
of the orthaceans rather than the porambonitaceans. It therefore seems reasonable to 
transfer the family to the Orthacea and to regard the stocks assigned to it as bearing 
the same relationship to the Orthidae as the Platystrophiinae do to the Plectorthidae 
and Finkelnburgiidae. 



ASTRABORTHIS gen. nov. 

Name. An orthacean shaped like a saddle (Gk. aoTpafSri) . 

Diagnosis. Subcircular, biconvex, uniplicate shells with a deeply sulcate pedicle 
valve ornamented by coarse angular costellae ; ventral interarea curved apsacline 
with wide, open delthyrium ; dorsal interarea curved anacline with wide, open 
notothyrium ; shell probably impunctate. 

Ventral interior with simple rounded teeth ankylosed to a pseudospondylium by 
solid deposits of secondary shell in adult valves, which may mask short receding 
dental plates, pedicle callist absent ; broad, medially divided adductor scars and 



ORDOVICIAN BRACHIOPODA 



69 




Fig. 6. Diagrammatic views of (A) the ventral and (B) the dorsal interiors of 

Astraborthis. 



submedial lobes of diductor scars impressed on the floor of a strong pseudospondy- 
lium ; mantle canal pattern saccate. 

Dorsal interior with a simple plate-like cardinal process and short, blade-like 
outward-curving brachiophores embedded in a thick deposit of secondary shell ; 
sockets oblique, notothyrial platform weak, fused with median ridge ; adductor 
scars quadripartite with smaller posterior pair ; mantle canal pattern probably 
digitate. 

Type species. Astraborthis uniplicata sp. nov. from the Mytton Flags. 

Discussion. Although the new genus is obviously alimbellid in its general 
morphology, it differs from other members of the family in its ornament and certain 
aspects of its internal morphology. Externally it is immediately distinguishable 
from the smooth Alimbella and finely capillate Medesia in its coarse costellation and 
internally in the weak development of the notothyrial platform and the blade-like 
nature of the brachiophores. It further differs from Alimbella in its wide ventral 
adductor scar and from Medesia in the absence of branching and medial fusion of 
the proximal parts of the vascula media. 

It may be of familial significance that dental plates have not yet been seen in 
adult pedicle valves of any of the three genera assigned to the Alimbellidae. It is, 
however, possible that short receding dental plates were developed during early 
growth stages but were later entirely buried in thick deposits of secondary shell 
secreted in the delthyrial cavity during formation of the pseudospondylium. Until 
young shells or their impressions have been examined it is safer to delay assessing 
the importance of this aspect of alimbellid morphology. 



70 SHELVE DISTRICT 

Astraborthis uniplicata gen. et sp. nov. 
(PI. ii, figs. 5, 6, 9, ii, 14 ; Text-fig. 6) 

Diagnosis. Subcircular, dorsibiconvex Astraborthis with the pedicle valve over 
four-fifths as long as wide and about one-third as deep as long ; ventral sulcus 
rounded and about two-thirds as wide as the valve length, ornamented by angular 
costellae with a wavelength of about 1 mm, 10 mm anterior of the ventral umbo with 
4 on the fold and up to 16 on the shell ; pseudospondylium about one-third as long 
as the pedicle valve. 

Description. Subcircular, strongly biconvex Astraborthis with the pedicle 
valve over four-fifths as long as wide and about one-third as deep as long ; ventral 
sulcus flanked by evenly convex lateral slopes, rounded in profile, about two-thirds 
as wide as the length of the pedicle valve, projecting antero-dorsally to fit into 
rounded dorsal fold ; brachial valve over one-third as deep as long, lateral slopes 
strongly and evenly convex ; ventral interarea over one-quarter as long as the pedicle 
valve, dorsal interarea about one-eighth as long as the brachial valve, cardinal 
extremities obtusely rounded ; radial ornamentation coarsely costellate with delayed 
costae originating at about 2 mm anterior of the umbo, with four occupying the crest 
of the fold, ribs sharply angular with a wavelength of about i-o mm at 10 mm 
anterior of ventral umbo. 

Ventral interior with pseudospondylium elevated on a solid platform of secondary 
shell, about three-quarters as wide as long, lateral boundaries subparallel, anterior 
boundary rounded extending forward from the ventral umbo for one-third the length 
of the pedicle valve. 

Dorsal interior with plate-like cardinal process obtusely triangular in lateral view, 
brachiophores short, curved, weakly divergent, adductor field strongly impressed 
divided by moderately strong median ridge and bounded posteriorly by weakly 
defined notothyrial platform. 

Type material. 

length width (mm) 
Holotype Incomplete external and internal moulds of 

pedicle valve (BB 35324a, b) 10-5 12-5 

Paratype Incomplete external and internal moulds of 

brachial valve (BB 35325a, b) - 

Horizon and locality. Mytton Flags exposed in small pathside quarry in the 
north-trending valley, 160 yds north-east of the north-east corner of Snailbeach 
Reservoir (Grid Ref. SJ 378024). 

Discussion. Only the incomplete moulds of a brachial and a pedicle valve, 
possibly the disarticulated remains of one shell, have so far been recovered from the 
Mytton Flags but the generic and specific individuality of the specimens cannot be 
doubted. Further collecting is desirable not only to assess the variation that 
occurred within the species but also to determine whether dental plates were developed 
in young specimens and to resolve the complete ribbing pattern, which, judged on 



ORDOVICIAN BRACHIOPODA 71 

the fragment of the dorsal exterior available at present, may have included about 
16 costae, arising just anterior of the umbo. 

Family FINKELNBURGIIDAE Schuchert & Cooper 1931 

DIPARELASMA Ulrich & Cooper 1936 

Diparelasma sp. 

(PI. 11, figs. 3, 7, 8) 

The internal and an incomplete external mould of two brachial valves, 2 and 3 mm 
long respectively, and an incomplete external mould of a pedicle valve, collected from 
the Mytton Flags are best assigned to the finkelnburgiid Diparelasma. The moulds of 
one brachial valve and the pedicle valve (BB 35335a, b) came from a small exposure, 
20 yds west of a runnel in the valley side north of the col separating Perkin's Beach 
and Mytton's Beach (Grid Ref. SJ 366002) ; those of the other brachial valve 
(BB 35336a, b) were obtained from outcrops near the top of the steep tributary to 
Crowsnest Dingle, 350 yds WNW of Blakemoorflat (Grid Ref. SJ 373008). 

The moulds indicate that the shell was ventribiconvex and transversely sub- 
quadrate in outline with obtuse cardinal angles, had a sulcate brachial valve about 
22% as deep as long and 64% as long as wide, and a pedicle valve about one-third 
as deep as long. Radial ornamentation was finely multicostellate with counts of 7 
and 9 costellae per mm, 2 mm anterior of the umbones. The cardinalia consisted 
of blade-like brachiophores diverging from the anacline interarea with bases con- 
verging onto the floor of the valve on either side of a low median ridge to define a 
deep notothyrium containing a simple cardinal process. The fulcral plates were 
more strongly developed in the smaller valve as antero-lateral boundaries to the 
sockets. A pair of discrete adductor muscle scars extending anteriorly of the umbo 
for about half the length of the valve were impressed on either side of the median 
ridge. 

Despite the lack of information about the ventral interior, the distinctive radial 
ornamentation and cardinalia leave little doubt that the moulds belong to Dipare- 
lasma. A number of species have been described from the Lower Ordovician, 
especially the Upper Canadian of N. America (Ulrich & Cooper 1938 : 147-156) ; but 
the cardinal process and dorsal sulcus are more strongly developed in the Shropshire 
specimens and may eventually prove to be diagnostic of a new species. 

Family PLECTORTHIDAE Schuchert & Le Vene 1929 

Subfamily PLECTORTHINAE Schuchert & Le Vene 1929 

PLECTORTHIS Hall & Clarke 1892 

Plectorthis whitteryensis sp. nov. 

(PI. 11, figs. 10, 13, 16, 17 ; PI. 12, fig. 1) 

Diagnosis. Semi-elliptical, slightly ventribiconvex, rectimarginate Plectorthis 
with a pedicle valve 69% as long as wide and 22% as deep as long, ornamented by 



72 SHELVE DISTRICT 

up to 27 costellae numbering 2 per mm, 5 mm antero-medially of the umbo, and 
bearing hollow embayments along the crests ; elongately oval ventral muscle scar 
extending forward for 28% of the length of the pedicle valve ; cardinal process 
ridge-like, simple. 

Description. Semi-elliptical, subequally biconvex Plectorthis, 3 pedicle valves 
averaging 69% as long as wide (range 64% to 73%) and 22% as deep as long (range 
21% to 24%) tending to become flatter away from the greatest depth of the valve 
in the umbonal region, and an evenly convex brachial valve with a mean depth 
relative to length of 19% (range 16% to 22%) for 3 valves ; anterior commissure 
rectimarginate, cardinal angles slightly obtuse ; radial ornamentation evenly cos- 
tellate with up to 27 flat-crested, parallel-sided costellae numbering 2 per mm, 5 mm 
antero-medially of the umbo and arising by dichotomy of primary costae within 
0-5 mm of the umbo, costellae becoming hollow at intervals of about 0-5 mm, 
interspaces finely striate and imbricate ; ventral interarea apsacline, delthyrium 
open, narrow, dorsal interarea shorter, anacline ; notothyrium open. 

Ventral interior with elongately oval muscle field averaging 74% (range 63% to 
93%) as wide as long and 28% (no range) as long as the length of the valve for 3 
pedicle valves, with well-developed submedial diductor lobes meeting antero- 
medially to enclose a lanceolate undifferentiated adductor scar ; teeth small, 
supported by short subparallel to narrowly divergent dental plates extending 
anteriorly for an average of 9-5% the length of the valve for 2 valves. 

Dorsal interior with a simple ridge-like cardinal process, brachiophores short, 
divergent, with bases slightly convergent to the floor of the valve and extending 
anteriorly for an average of 10% of the length for 3 valves ; short fulcral plates 
defining small suboval sockets, median ridge absent, adductor scars only rarely 
impressed as a pair of widely spaced bilobed scars extending anteriorly for about 
two-fifths the length of the brachial valve ; mantle canal systems unknown. 
Type specimens. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 35462a, b) 6-5 io-o 

Paratypes External and internal moulds of brachial valve 

(BB 35463a, b) 5-0 

External and internal moulds of brachial valve 

(BB 35464a, b) 5-0 8-5 

External and internal moulds of pedicle valve 

(BB 37158a, b) 6-o 9-2 

Horizon and localities. Whittery Shales exposed in the lower part of the 
west bank of the River Camlad 60 yds north-east of Marrington Farm (Grid Ref. 
SO 272970). 

Discussion. Many Ordovician Plectorthis have been described, especially from 
North America (see Cooper 1956 : 447-456), but the species collected from the 
Whittery Shales is unique in the combination of a number of distinctive features, 
in particular the semi-elliptical outline and the peripheral flattening of the pedicle 



ORDOVICIAN BRACHIOPODA 73 

valve, the absence of a dorsal sulcus, the relative fineness of the costellate ornamenta- 
tion, the variable development of a simple cardinal process and the occurrence of 
hollow embayments along the crests of the costae. Indeed, in respect of the last 
two features, the new species shows some affinity with Desmorthis Ulrich & Cooper 
1936, and in view of the essential homogeneity of the American Plectorthis which 
suggests that they constitute a closely related species group, the Shropshire specimens 
may represent a stock with only remote ancestral links with their American con- 
temporaries. 

Plectorthis sp. 
(PI. 11, figs. 12, 15) 

A complete internal mould and a fragment of the external mould (BB 35466a, b) of 
an orthacean pedicle valve, collected from Aldress Shales exposed in a bank of Ox 
Wood Dingle at the south-west corner of Ox Wood, a few yards north of the Rorring- 
ton-Wotherton road (Grid Ref. SJ 290007), have been provisionally identified as 
Plectorthis. The valve, which was about 13 mm long and 14-5 mm wide, was subcircular 
in outline with obtusely rounded cardinal angles and just over one-sixth as deep as 
long becoming flatter antero-medially. The surface was ornamented by micro- 
scopic concentric lamellae and 24 ribs which must have arisen at or near the umbo 
except for a few secondary costellae in the lateral areas. The ribs were rounded 
with a wavelength of o«6 mm, 10 mm anterior of the umbo, and were separated by 
interspaces about 0-5 mm wide. The interarea was short, curved, and apsacline, 
the delthyrium open, and the small teeth were supported by widely divergent dental 
plates extending anteriorly for less than one-sixth the length of the valve. No 
identifiable muscle scars were impressed on the floor of the valve. 

In the absence of impressions of the brachial valve, no valid comparisons can be 
drawn with described Plectorthis. The pedicle valve, however, is immediately 
distinguishable from that of P. whitteryensis sp. nov. in its subcircular rather than 
semi-elliptical outline, its significantly fewer ribs which are, moreover, rounded not 
flattened, and its widely divergent dental plates. 



DESMORTHIS Ulrich & Cooper 1936 

Desmorthis ? sp. nov. 

(PI. 12, figs. 2-4, 8) 

Incomplete external and internal moulds of a brachial valve (BB 35337a, b), from 
exposures in a runnel to the east of the footpath to Blakemoorflat at the head of 
Mytton's Beach (Grid Ref. SJ 373006), are the sole representatives in the Mytton 
Flags of a highly distinctive plectorthid genus. The valve which was 7 mm long, 
semi-oval in outline, rectimarginate and gently convex medially with flattened 
lateral slopes, was ornamented by strong, sharply rounded hollow costellae number- 
ing 2 per mm, 5 mm antero-medially of the umbo, and a subdued concentric orna- 
mentation of faint lamellae. The planar anacline interarea was about one-tenth as 



74 SHELVE DISTRICT 

long as the valve, the cardinal process a fine median ridge on a low notothyrial plat- 
form, separating two subtriangular diductor impressions. The short divergent 
brachiophores were pointed with low convergent bases bounding the notothyrial 
platform, while the narrow sockets were defined by well-developed fulcral plates. A 
linear median ridge extended anteriorly of the notothyrial platform for about two- 
fifths of the valve length, between suboval, poorly impressed adductor scars. 

Two sets of external and internal moulds (BB 35338a, b ; BB 35339) from the 
Stapeley Volcanic Group collected from outcrops in an adit near Perkin's Beach, 
west of Shepherd's Rock, represent pedicle valves which appear to be conspecific 
with the brachial valve described above. The ornamentation, consisting of faint 
concentric lamellae and rather widely separated sharply rounded hollow costellae, 
is identical. Unfortunately little can be said about the internal morphology. In 
the larger specimen, which is 4 mm long, the ventral muscle field is not impressed 
and only the mould of a divergent dental plate, extending forward for about one- 
seventh the length of the mould, is seen. 

The profiles of the valves, the development of hollow costellae and, with regard to 
the brachial valve from the Mytton Flags, the nature of the cardinalia are all 
indicative of the plectorthid affinities of the specimens. The simplicity of the car- 
dinal process further suggests that the brachial valve, at least, is closer to Desmorthis 
than any other genus assigned to the family. However, the convergence of the 
brachiophores and the lamellose exterior hint that when the species is fully known it 
will prove to be generically distinct from other plectorthids. 

GELIDORTHIS Havlicek 

Gelidorthis cf. partita (Barrande) 

(PI. 12, figs. 5-7) 

1879 Orthis partita Barrande : plate 63, fig. II. 

1950 Giraldiella partita (Barrande, 1879) Havlicek : 124. 

1968 Gelidorthis partita (Barrande, 1879) Havlicek : 125. 

1971 Gelidorthis partita (Barrande, 1879) Havlicek : 42. 

A small number of impunctate valves never more than 8 mm long, from the 
Rorrington Beds exposed in Spy Wood Brook 20 yds downstream from where it 
joins Deadman's Dingle (Grid Ref. SO 289958), most closely compare with Gelidorthis 
partita (Barrande) from the Lower and Middle Caradoc rocks of Czechoslovakia. 
The shells were subquadrate in outline with an evenly convex pedicle valve 
(BB 37159 ; BB 35418a, b) about three-quarters as long as wide and about one- 
eighth as deep as long, and a very gently convex brachial valve (BB 35419a, b) 
lacking a definite sulcus. The relatively coarse radial ornamentation consists of 
angular costellae numbering 3 per mm at 5 mm antero-medially of the ventral umbo. 
The ribbing is strongly impressed internally even in the ventral umbonal cavity 
which is defined by short, widely divergent dental plates extending forward for about 
one-tenth the length of the pedicle valve. As in the pedicle valve, the dorsal muscle 
field is also obscured by internal ribbing ; but the cardinalia are well developed and 
consist of a cardinal process incipiently differentiated into a myophore and shaft, 



ORDOVICIAN BRACHIOPODA 75 

brachiophores supported by short convergent bases, and sockets denned by small 
fulcral plates. 

These features are consistent with the variability described by Havlicek (1950 : 
124 ; 1971 : 42) in young specimens of G. partita, although the brachial valve of the 
Shropshire Gelidorthis appears to be relatively flatter. However only two of these 
valves are known at present and until more material including adult specimens is 
available for the significance of this difference to be assessed, there is no justification 
in according them systematic recognition. 



TAZZARINIA Havlicek 1971 

Tazzarinia elongata sp. nov. 

(PI. 12, figs. 9-12, 15) 

Diagnosis. Elongately subquadrate, ventribiconvex Tazzarinia with a coarsely 
fascicostellate ornamentation and a dorsal sulcus dying out anteriorly ; bilobed 
ventral muscle scar extending forward for almost 30% the length of the pedicle 
valve ; pointed brachiophores supported by short bases. 

Description. Subquadrate, ventribiconvex Tazzarinia with slightly obtuse 
cardinal angles ; pedicle valve almost as long as wide and one-quarter as deep as 
long, slightly carinate medially with evenly convex lateral slopes ; brachial valve 
about four-fifths as long as wide and one-sixth as deep as long, with a narrow median 
sulcus becoming indistinct beyond the 4 mm growth stage, lateral slopes evenly 
convex ; radial ornamentation strongly fascicostellate with narrowly rounded 
costellae numbering 5 per mm, 3-5 mm antero-medially of the dorsal umbo ; ventral 
interarea slightly curved apsacline, longer than the anacline dorsal interarea, 
notothyrium and delthyrium open but pedicle callist strongly developed within the 
latter. 

Teeth supported by narrowly divergent dental plates extending forward for about 
one-sixth the length of the pedicle valve ; ventral muscle scar strongly bilobed with 
a median adductor track and flanking diductors impressed anteriorly for almost 
three-tenths the length of the valve. 

Cardinal process massive with crenulated posterior face, notothyrial platform well 
developed passing anteriorly into median ridge and flanked by convergent bases of 
pointed brachiophores, sockets elongate with small fulcral plates ; suboval adductor 
scars impressed on either side of median ridge and extending anteriorly for over 
two-fifths the length of the brachial valve. 

Type material. 

length width (mm) 
Holotype External and internal moulds of brachial valve 

(BB 35332a, b) 5-5 

Paratype External and internal moulds of pedicle valve 

(BB 35333a, b) 3-8 3-8 



76 SHELVE DISTRICT 

Type horizon and locality. Meadowtown Beds exposed in a small excavation 
in the corner of a field alongside the cart-track from Meadowtown to Waitchley, 
330 yds south of Waitchley (Grid Ref. SJ 311017). 

Discussion. Although the moulds of only one pedicle and one brachial valve of 
Tazzarinia have been recovered from the Meadowtown Beds, there is no doubt about 
their generic affiliations or their specific distinctiveness. The specimens compare 
in every important respect with the Tazzarinia described by Havlicek (1971 : 39) 
from the Lower Caradoc of Morocco except that bilobation of the cardinal process 
cannot be confirmed in the Shropshire dorsal mould. Yet the Shelve form is im- 
mediately distinguishable from both Moroccan species, T. drotae and T. foraminosa, 
especially in its coarsely fascicostellate ornamentation and its elongately subquadrate 
outline. 



Subfamily PLATYSTROPHIINAE Schuchert & Le Vene 1929 

PLATYSTROPHIA King 1850 

Platystrophia caelata sp. nov. 

(PI. 12, figs. 13, 14, 16-19) 

Diagnosis. Subquadrate, dorsibiconvex Platystrophia with the pedicle valve 
76% as long as wide and 32% as deep as long, with a flat-bottomed sulcus 54% as 
wide as the valve length, ornamented by 2 costae in the sulcus and 10 to 14 on the 
lateral slopes with lamellae extending along the crests of costae as short rounded 
canopies. 

Description. Subquadrate, dorsibiconvex, uniplicate Platystrophia with sub- 
rectangular cardinal angles, the brachial valve just over two-fifths as deep as long, 
the pedicle valve on average 76% (range 70% to 79%) as long as wide and 32% 
(range 28% to 35%) as deep as long for 3 valves, with a flat-bottomed sulcus with a 
mean width of 54% (range 50% to 57%) relative to the length for 3 valves, and 
evenly convex lateral slopes ; dorsal interarea short, anacline, ventral interarea 
curved, apsacline about one-fifth as long as pedicle valve ; costae angular with a 
wavelength of just over 1 mm, 5 mm anterior of dorsal umbo, covered with densely 
distributed tubercles and numbering 2 in the ventral sulcus with 10 to 14 on the 
lateral slopes ; lamellae strongly developed and, beyond 4 mm from umbones, 
extending along the crests of the costae as short, rounded canopies partially covering 
subtriangular concave pads of secondary shell. 

Ventral interior with small pedicle callist, thick teeth supported by short, receding 
dental plates extending anteriorly for less than one-fifth the length of the pedicle 
valve, and subtriangular muscle field with wide adductor impressions occupying the 
floor of the valve, less than half as wide as long and extending anteriorly for just 
over two-fifths the length of the pedicle valve ; mantle canal impressions obscure 
but ventral system possibly digitate with abbreviated vascula media occupying the 



ORDOVICIAN BRACHIOPODA 77 

crests of the costae in the sulcus and gonocoels spreading antero-laterally along the 
crests of the lateral costae. 

Dorsal interior with ridge-like cardinal process and wide notothyrium bounded by 
secondarily thickened short brachiophores extending anteriorly for about one-fifth 
the length of the brachial valve ; dorsal adductor impressions quadripartite, extend- 
ing anteriorly for almost half the length of the brachial valve on either side of a low 
median ridge with the subtriangular anterior pair the larger. 

Type material. 

length width (mm) 
Holotype External and internal moulds of brachial valve 

(BB 35493a, b) 9-5 

Paratypes External and internal moulds of pedicle valve 

(BB 35494a, b) 9-0 117 

External and internal moulds of pedicle valve 

(BB 35495a, b) 10-5 15-0 

Horizon and localities. Whittery Shales : BB 35493 from Whittery Quarry 
at the south end of Whittery Wood, near Chirbury (Grid Ref . SO 275981) ; BB 35494, 
BB 35495 from exposures in the lower part of the west bank of the River Camlad, 
60 yds north-east of Marrington Farm (Grid Ref. SO 272970). 

Discussion. The Platystrophia which occurs rarely in the Whittery Shales has a 
highly distinctive concentric ornamentation of differentially developed lamellae. 
Most Platystrophia are variably lamellose and some such as P. chama Eichwald (as 
figured by Alichova 1953 : 131) strikingly so, but in P. caelata sp. nov. more than 
4 mm long strong lamellae occur at regular intervals of about 0-5 mm and their 
forward growth along the costae was accompanied by an equally regular sag of the 
mantle edge. The resultant pattern is a series of subcircular, narrow canopies, 
aligned along the crests of the costae, inclined at about 35 ° to the shell surface and 
overhanging the posterior parts of alternating, depressed, subtriangular areas of 
secondary shell. In the intercostate spaces, the lamellae lie flat on one another as a 
series of overlapping tongues directed anteriorly. This ornament is characteristic 
of all eight specimens of the genus collected from the Whittery Shales. It is un- 
known in any other described Platystrophia, and in recognition of its uniqueness a 
new species has been erected to include the Whittery shells. 

In other features, the new species appears to be unexceptional, although this may 
be due to the meagreness of the sample. Judging from the clarity of gonadal im- 
pressions and muscle scars which are deeply inserted even on the notothyrial floor 
and about the cardinal process, specimens attaining a length of 10 mm were gerontic 
individuals so that at least the Whittery representatives of the species were small in 
size. All specimens showing the disposition of the costae on the fold and sulcus 
were biplicate. The interiors of both valves show less secondary thickening than is 
typical of later species of Platystrophia, but, even so, the brachiophore bases in the 
brachial valve are heavily encased in secondary shell as are the ventral umbonal 
chambers. 



78 SHELVE DISTRICT 

Platystrophia cf. major Williams 1955 
(PL 13, figs. 1, 2, 4) 

1955 Platystrophia precedens McEwan major Williams : 402. 

The moulds of two immature valves of 'Platystrophia lacking the distinctive 
concentric ornamentation of P. caelata sp. nov. have also been found in the Shelve 
area. They are the moulds of a brachial valve nearly 5 mm long (BB 35586a, b) 
from Aldress Shales exposed in the bank of Ox Wood Dingle at the south-west corner 
of Ox Wood a few yards north of the Rorrington-Wotherton road (Grid Ref. 
SJ 290007), and a pedicle valve 5-5 mm long (BB 35587a, b) from the Spy Wood 
Grit exposed on top of the ridge 1440 yds NNE of the Rorrington bench mark 599 
(Grid Ref. SJ 303018). Neither valve can be completely reconstructed from the 
indifferently preserved moulds but the brachial valve appears to have been about 
80% as long as wide and 20% as deep as long with a well-defined rounded fold about 
40% as wide as the valve. The radial ornamentation consists of 3 costae on the 
fold and 5 on each lateral slope with a wavelength of 07 mm, 5 mm antero-medially 
of the umbo. Only impressions of subparallel brachiophore bases and a blade-like 
cardinal process are preserved internally. The pedicle valve was similarly orna- 
mented, although traces of fine concentric lamellae are also sporadically preserved, 
while internally the ventral muscle scar extended well beyond the slightly divergent 
dental plates for 36% of the valve length. 

These few statistics and especially the presence of 3 costae on the fold and 2 in 
the sulcus suggest that both valves may be provisionally assigned to P. major 
Williams (in Whittington & Williams 1955 : 402) first described from the Derfel 
Limestone of N. Wales. The species is morphologically like the American species 
P. precedens McEwan but is now also known from the Upper Llandeilo (MacGregor 
1961 : 184). New evidence on intraspecific variability in the arrangement of costae 
suggests that the species is more likely to be related to the Platystrophia dentata 
(Pander) group from the Baltic (Williams 1963 : 371). 



MCEWANELLA Foerste 1920 
Mcewanella sp. 

(PL 13, figs. 3, 5) 

An incomplete pedicle valve (BB 35416) and a deformed internal ventral mould 
(BB 35417) from the Whittery Shales exposed in the stream at the north end of 
Spring Coppice 865 yds south-east of Hockleton Bridge (Grid Ref. SO 279997) are 
the only specimens of Mcewanella so far recovered in the Shelve area. 

Growth lines on the valve which was about half as deep as long, and shallowly 
sulcate antero-medially, indicate that the outline changed from semi-circular to 
subquadrate during growth, because the immature valve was widest along the 
mucronate hinge-line whereas fully developed valves were about as wide as long with 
the maximum width well anterior of the hinge line with its orthogonal cardinal 
angles. The external surface was ornamented by branching costellae numbering 



ORDOVICIAN BRACHIOPODA 79 

3 per mm, 5 mm antero-medially of the umbo and further corrugated by up to 7 
radiating rounded costae with a wavelength of about 3 mm, 10 mm antero-medially 
of the umbo. 

The ventral mould is distorted umbonally but there is evidence of an apsacline 
interarea and massive trigonal teeth supported by strong, narrowly divergent dental 
plates. The ventral muscle scar was elongately oval with a broad median adductor 
track reaching to the anterior margin. 

The specimens are closely related to the Upper Llandeilo M. berwynensis (Mac- 
Gregor 1961 : 183). However, the Welsh Mcewanella bears a ventral fold, and 9 
superimposed costae, and until the variability of these features can be ascertained 
in the Shropshire form it seems safer to withhold specific identification. 



SALACORTHIS gen. nov. 

Name. An orthacean with an external ornament like a sieve (Gk. adXai). 

Diagnosis. Subquadrate, biconvex, uniplicate shells with a strong dorsal median 
fold emerging from a neanic sulcus and a complementary ventral sulcus ; radial 
ornamentation coarsely costate to costellate with two or three thick rounded costae 
on the fold and up to three equally wide costae on each lateral slope, costellae arising 
sporadically in later growth stages either by division of costae or as intercalations in 
the interspaces ; external shell surface also ornamented by densely distributed, 
deep exopuncta arranged qnincnncially ; ventral interarea curved, apsacline with 
open triangular delthyrium, dorsal interarea shorter, slightly curved anacline, noto- 
thyrium open ; shell impunctate. 

Ventral interior with strong teeth supported by short but well-defined dental 
plates extended on the valve floor as raised ridges laterally bounding an elongately 
oval undifferentiated muscle field ; ventral mantle canal system unknown. 

Dorsal interior with a wide notothyrial platform bearing a median ridge-like 
cardinal process, brachiophores rod-like, divergent, continuous laterally with strong 
concave fulcral plates defining oblique, hemiconical sockets ; brachiophore bases 
large, convergent onto notothyrial platform and median ridge ; dorsal adductor 
scars divided by median ridge into a pair of triangular undifferentiated impressions 
with the apices directed posteriorly and deeply inserted beneath the anterior borders 
of the brachiophore bases ; dorsal mantle canal system unknown. 

Type species. Salacorthis costellata sp. nov. from the Spy Wood Grit. 

Discussion. The new genus, with the exception of a poorly preserved exfoliated 
pedicle valve revealing the impunctate nature of the shell, is represented solely by a 
few moulds from the Spy Wood Grit. Yet its affinities are not in doubt because it 
combines the typical cardinalia of the Plectorthidae with the coarsely costate 
ornamentation and strongly uniplicate anterior margin of the Platystrophinae. 
Indeed with regard to its radial ornamentation it anticipates the relationship between 
Platystrophia and Mcewanella (Schuchert & Cooper 1932 : 69). The former is costate 
and the latter is believed, at least in respect of the American species, to have been 



80 SHELVE DISTRICT 

derived from Platystrophia by the superimposition of a multicostellate ornament. 
The beginnings of a similar development can be seen in one external mould of 
Salacorthis (PI. 13, fig. 12) but it was not necessarily typical of the species because 
no costellae were developed in a larger pedicle valve and even in the figured specimen 
costellae did not appear until the valve was about 7 mm long. The new genus 
therefore differs from both Platystrophia and Mcewanella in being sporadically 
costellate. 

The principal difference, however, and one which immediately distinguishes 
Salacorthis from the other two members of the Platystrophinae, lies in the con- 
spicuously exopunctate appearance of the shell surface. The exopuncta were deep 
coarse pits penetrating the external shell vertically for depths of about o-i mm and 
arranged in a closely crowded quincuncial pattern. They are quite different in 
morphology and origin from the obliquely disposed apertures which constitute the 
'hollow ribs' so typical of the plectorthids (Williams in Williams et al. 1965 : H70), 
and are all the more unexpected because the external surface of Platystrophia is 
noteworthy among orthaceans for being distinctively pustulose. 

Internal characters that indicate the platystrophiinid affinities of Salacorthis 
include the simple cardinal process and the suboval ventral field which, although 
undifferentiated in known specimens of the new genus, probably consisted of a broad 
adductor scar flanked by elongate bases for the diductors and adjustors, as in 
Platystrophia. The brachiophores and their supporting bases are also comparable, 
although the degree to which the posterior parts of the dorsal adductor impressions 
are inserted behind the brachiophore bases of Salacorthis is unmatched in Platy- 
strophia and Mcewanella. But such internal differences as exist are minor and do 
not affect the conclusion that Salacorthis arose out of the earlier established Platy- 
strophia. 



Salacorthis costellata gen. et sp. nov. 
(PI. 13, figs. 6-13 ; Text-fig. 7) 

Diagnosis. Dorsibiconvex, subquadrate Salacorthis with a brachial valve about 
three-quarters as long as wide and 37% as deep as long ; dorsal fold almost half as 
wide as valve length, bearing 2 or 3 costae with a wavelength of 1-5 mm, 5 mm 
antero-medially of the umbo, with 2 or 3 on the flanks, all becoming costellate in 
late growth stages ; ventral muscle scar suboval extending forward for nearly one- 
third the length of the pedicle valve ; dorsal median septum short. 

Description. Dorsibiconvex, subquadrate Salacorthis with cardinal angles 
becoming obtuse in adult growth stages, pedicle valve about one-quarter as deep 
as long, brachial valve three-quarters as long as wide with a mean depth relative to 
length of 37% for 3 brachial valves ; dorsal fold 49% as wide as the length of 3 
brachial valves bearing 2 or 3 strong costae with a wavelength of about 1-5 mm, 
5 mm antero-medially of the dorsal umbo, with 2 or 3 additional costae on each of 
the lateral slopes, all costae may split into costellae in late growth stages ; exopuncta 
coarse with counts of 3 per 0-5 mm, 5 mm anterior of the umbo. 



ORDOVICIAN BRACHIOPODA 



81 




cardinal process 
socket 
— tooth 



brachiophore 



dental plate 



muscle scar 




Fig. 7. Diagrammatic views of (A) the ventral and (B) the dorsal interiors of 

Salacorthis. 



Ventral interior with subparallel dental plates extending forward for one-seventh 
the pedicle valve length, suboval ventral muscle scar undifferentiated, extending 
anteriorly for nearly one-third the length of the pedicle valve. 

Dorsal interior with brachiophore bases extending forward for over one-quarter 
the length of the brachial valve ; median septum supporting the cardinalia short, 
extending anteriorly for two-fifths the length of the brachial valve. 

Type material. 



Holotype 



length width (mm) 



Paratypes 



37 
ii-o 



5-o 

12-0 



4-5 



External and internal moulds of brachial valve 

(BB 37153a, b) 

Internal mould of pedicle valve (BB 37154) 

Incomplete external mould of brachial valve 

(BB 37155) 

Incomplete external mould of brachial valve 

(BB 37156) 

Incomplete external mould of brachial valve 

(BB 37157) 

Horizon and localities. Spy Wood Grit : BB 37153, 37154 from exposures on 
top of ridge 1440 yds NNE of Rorrington bench mark 599 (Grid Ref. SJ 303018) ; 
BB 37155, 37157 from outcrops 1100 yds NNE of Rorrington (Grid Ref. SJ 303015) ; 
BB 37156 from outcrops in the north bank of Spy Wood Brook 170 yds north-east 
of Spy Wood Cottage (Grid Ref. SO 282958). 

Discussion. No other species congeneric with Salacorthis costellata has yet been 
so described although in view of the strongly developed platystrophiinid characters 
of the Spy Wood form and the microscopic nature of its chief differentiating feature, 
some stocks, currently recorded as Platystrophia or Mcewanella, may ultimately 
prove to belong to the genus. 



82 SHELVE DISTRICT 

Family SKENIDIIDAE Kozlowski 1929 

Genus SKENIDIOIDES Schuchert & Cooper 1931 

Skenidioides cf. costatus Cooper 

(PI. 13, figs. 14-16 ; PI. 14, figs. 1-3) 

1956 Skenidioides costatus Cooper : 493. 

1963 Skenidioides cf. costatus Cooper; Williams : 375. 

Diagnosis. Subpyramidal, ventribiconvex Skenidioides with a carinate pedicle 
valve 54% as long as wide and 55% as deep as long, ornamented by up to 8 costae 
on each of the lateral slopes and a thick median costa giving rise to a pair of lateral 
costellae 1*5 mm anterior of the umbo. 

Description. Semi-oval, sulcate Skenidioides with a subpyramidal, carinate 
pedicle valve with a mean length relative to width of 54% (range 50% to 59%) for 
4 valves and a mean depth relative to length of 55% (range 53% to 58%) for 3 valves ; 
brachial valve very gently convex with a sharp median sulcus ; radial ornamentation 
costate with up to 8 rounded costae occupying each lateral slope but with ventral 
carina defined by thick rounded costa, about 0-5 mm in wavelength 2 mm anterior 
of ventral umbo, giving rise to a pair of lateral costellae about 1-5 mm anterior of 
ventral umbo. 

Ventral interior with free spondylium about one-fifth as deep as the length of the 
valve. 

Dorsal interior with thin and plate-like cardinal process continuous with high 
median septum, brachiophore bases converging onto dorsal septum at an average of 
33% the length of 2 brachial valves ; dorsal adductor field as a pair of subtriangular 
impressions divided by median septum and extending anteriorly for nearly three- 
fifths the length of the brachial valve. 

Figured material. 

length width (mm) 

Internal mould of brachial valve (BB 35409) 2-3 4-4 

Incomplete internal mould of pedicle valve (BB 35410) 2-5 4-4 

Internal mould of pedicle valve (BB 3541 1) 1-2 27 

Internal mould of pedicle valve (BB 35412) 0-8 1-5 

Incomplete external mould of pedicle valve (BB 35413) 1*5 - 

Horizons and localities. Spy Wood Grit : BB 35411 to 35413 from exposures 
in north bank of Spy Wood Brook, 170 yds north-east of Spy Wood Cottage 
(Grid Ref. SO 282958). Whittery Shales : BB 35409 from exposures in the lower 
part of the west bank of the River Camlad, 60 yds north-east of Marrington Farm 
(Grid Ref. SO 272970) ; BB 35410 from exposures in the stream at the north end 
of Spring Coppice, 865 yds south-east of Hickleton Bridge (Grid Ref. SO 279997). 

Discussion. Moulds of Skenidioides occur only rarely in both the Spy Wood 
Grit and Whittery Shales but are clearly related to the species found in the Gelli-grin 



ORDOVICIAN BRACHIOPODA 83 

Calcareous Ashes which has been compared with S. costatus Cooper from the Porter- 
field of North America (Williams 1963 : 375). In internal proportions and the shape 
of the shell the likeness is certainly close and, although the Shelve pedicle valve 
tended to be relatively wider and deeper than the average Bala one, the differences 
are not important. Indeed the only indication of a difference which may eventually 
prove to be important is in the lack of development of a costellate ornamentation 
on the Shelve Skenidioides. The sample is indistinguishable from S. costatus in the 
number and wavelength of costae, but only the ventral median costa of the Shelve 
specimens gave rise to costellae, whereas external branches arose freely from the 
lateral costae of the Bala stock. This difference is not entirely related to the overall 
greater size attained by the Bala shells. Within the size range of the Shelve speci- 
mens, which are between 0-8 and 2-0 mm long, costellae branched from one or more 
lateral costae in 7 out of 12 pedicle valves from the Gelli-grin Group. It is therefore 
possible that the Shelve Skenidioides were characterized by a simplicity of ribbing 
that persisted throughout growth, but larger collections will have to be obtained to 
determine whether taxonomic recognition of the difference is worth while. 



PROTOSKENIDIOIDES gen. nov. 
Name. One of the first skenidiids. 

Diagnosis. Transversely semi-oval, plano-convex shells with a high carinate 
pedicle valve and sharply sulcate brachial valve ; radial ornamentation costate 
with a wide ventral median costa forming the carina of the pedicle valve and cor- 
responding to the dorsal sulcus ; ventral interarea long, planar, apsacline with an 
open delthyrium, dorsal interarea shorter, planar, anacline ; shell probably im- 
punctate. 

Ventral interior with pointed teeth and shallow spondylium apparently free of the 
valve floor ; muscle and mantle canal impressions unknown. 

Dorsal interior with divergent brachiophores supported by flat-lying brachiophore 
bases, discrete and transversely lobate in young shells but extending anteriorly as 
elevated plates in adult shells and converging medially to form a ridge-like cardinal 
process which is joined by a low swelling on the valve floor with a high long median 
septum ; sockets narrow oblique, defined by concave elevated fulcral plates in 
adult shells only ; adductor muscle field divided by median septum into two elongated 
oval impressions bounded laterally by raised curved ridges ; dorsal mantle canal 
system unknown. 

Type species. Protoskenidioides revelata sp. nov. from the Mytton Flags. 

Discussion. The skenidiid found in the Mytton Flags is represented exclusively 
by fragile moulds of what must have been thin-shelled valves, but although the 
smallest impression of the brachial valve recovered is about 4 times as long as that 
of the protegulum, the size range (o-8 mm to 2-0 mm in valve length) is sufficiently 
great to reveal some hitherto unsuspected aspects of the origin and development of 
the skenidiid cardinalia. 



8 4 




SHELVE DISTRICT 

cardinal process 



socket 
brachiophore 
brachiophore base 





median septum 





A 


B 


c 


0-0-9 


5 


1 




1-0-1-4 




2 




1-5-2-0 




1 


4 



Fig. 8. Diagrammatic views of the dorsal interiors of Protoskenidioides showing three 
stages (A, B, C) in the development of the cardinalia during ontogeny ; with a correla- 
tion table below indicating the relationship between the length of the brachial valve 
and the illustrated growth stages of the cardinalia. 



The smallest moulds show that when brachial valves were less than i mm long the 
brachiophores were well developed and widely divergent (Text-fig. 8). Postero- 
laterally they were flanked by simple slot-like sockets indented on the valve floor. 
Antero-medially they were continuous with a pair of transversely oval areas defined 
by slightly elevated anterior boundaries that converged towards the notothyrial 
apex of the valve. It is assumed that these wide, discrete areas represented the 
seats for the attachment of the dorsal ends of the diductor muscles rather than 
supports for the brachiophores although it is convenient to refer to them as brachio- 
phore bases. At this growth stage there was no notothyrial platform and the valve 
interior was dominated by the median elevation representing the external sulcus 
bearing a small, thin, median septum. 

With further growth the antero-lateral parts of the sockets were raised above the 
valve floor by the development of concave fulcral plates and the brachiophore bases 
were extended anteriorly as a pair of lobate platforms which were also free of the 
valve floor. The bases also encroached onto the median elevation which bore a 
small ridge in some valves, finally to converge and form a simple cardinal process 
clearly made up of the adjacent borders of the brachiophore bases and contained 
within a low, anteriorly bilobed septalium. Concomitantly the median septum 



ORDOVICIAN BRACHIOPODA 85 

extended anteriorly and also posteriorly almost to unite with the cardinal process 
from which it was separated by a low saddle representing a remnant of the median 
elevation. 

These modifications leading to the definition of the septalium are unknown in 
younger skenidiids and serve to distinguish Protoskenidioides from other members 
of the family. In S. cf. costatus Cooper from the Gelli-grin Group, for example, 
impressions of the smallest brachial valves known, which are just over 1 mm long, 
show that the septalium was not bilobed but evenly convex anteriorly as in adult 
valves, and that it contained a high cardinal process not yet differentiated into 
a myophore and shaft but which was continuous with the median septum and 
bore no sign of having been built up from the median edges of the brachiophore 
bases. 

The likelihood that Protoskenidioides gave rise to Skenidioides by an accelerated 
development of a well-defined septalium is strengthened by the close comparison 
between the two genera in other characteristics such as the distinctive shape and 
ornamentation. Even the inference that the spondylium in the pedicle valve of 
Protoskenidioides is free does not militate against the relationship because a septal 
support to the spondylium of Skenidioides cannot always be identified without 
recourse to apical sections of the shell. Such a line of descent is also still consistent 
with the suggestion (Kozlowski 1929 : 129) that the skenidiids were derived from the 
finkelnburgiids : indeed the attitude of the slightly inclined, convergent brachio- 
phore bases of Protoskenidioides is more reminiscent of Finkelnburgia than the later 
skenidiids. 



Protoskenidioides revelata gen. et sp. nov. 
(PI. 14, figs. 4-13 ; Text-fig. 8) 

Diagnosis. Subpyramidal, mucronate Protoskenidioides with a planar brachial 
valve 47% as long as wide, lateral slopes ornamented by up to 18 costae with a broad 
ventral median costa corresponding to a dorsal sulcus 0-4 mm wide, 1-5 mm antero- 
median^ of the umbo. 

Description. Transversely semi-oval, mucronate Protoskenidioides with a sub- 
pyramidal pedicle valve and a planar brachial valve averaging 47% (range 45% to 
48%) as long as wide for 4 valves with a well-defined rounded median sulcus con- 
sistently 0-4 mm wide, 1-5 mm antero-medially of the umbo of 4 valves; radial 
ornamentation on lateral slopes consisting of up to 18 costae increasing in number 
during growth by implantation with the right lateral consistently 0-2 mm wide in 
4 valves, and the ventral median costa complementary to the dorsal sulcus about 
0-6 mm wide in adult shells. 

Ventral interior with shallow spondylium, broadly convex anteriorly. 

In adult brachial valves the septalium, about half as long as wide, extended 
anteriorly for an average of 23% (range 21% to 26%) the length of 5 valves but it 
was not continuous with a long median septum which may have reached the anterior 



86 SHELVE DISTRICT 

margin, adductor muscle field extending anterior of dorsal umbo for almost three- 
quarters the length of the valve. 

Type material. 

length width (mm) 

Holotype Internal mould of brachial valve (BB 35387) 17 3-0 

Paratypes External and internal moulds of brachial valve 

(BB 35388a, b) 1-3 2-6 

Incomplete internal mould of pedicle valve 

(BB 35389) _ 

Incomplete internal mould of brachial valve 

(BB 35390) 1-2 

Incomplete external mould of brachial valve 

(BB 35391) 

Incomplete internal mould of brachial valve 

(BB 35392) o-8 

Incomplete internal mould of brachial valve 

(BB 35393) i-i 

External and internal moulds of brachial valve 

(BB 35394a, b) 1-6 

External and internal moulds of pedicle valve 

(BB 35395a, b) i-i 2-0 

Internal mould of brachial valve (BB 35396) 1-9 - 

External and internal moulds of brachial valve 

(BB 35397^, b) i-6 

External and internal moulds of brachial valve 

(BB 35398a, b) 1-5 

External and internal moulds of brachial valve 

(BB 35399a, b) 17 - 

External and internal moulds of brachial valve 

(BB 35400a, b) 2-0 

External and internal moulds of brachial valve 

(BB 35401a, b) 0-9 17 

Horizon and localities. Mytton Flags : BB 35387 to 35393, BB 35398, 35399 
from a trench excavated in Shelve Church Beds (Grid Ref. SO 335990) ; BB 35394 
to 35397 from exposures of Shelve Church Beds along a cart-track near Wood House, 
Gravels (Grid Ref. SJ 338003) ; BB 35400, 35401 from Shelve Church Beds exposed 
along a cart-track 50 yds north-west of Wood House, near Gravels (Grid Ref. 

SJ 337002). 

Discussion. The moulds of the new species that have been collected are so 
closely crowded and so commonly collapsed that only a few statistical data were 
obtainable to indicate the variability of the new species. These are included in the 
description given above. 



ORDOVICIAN BRACHIOPODA 87 

Superfamily ENTELETACEA Waagen 1884 

Family SCHIZOPHORIIDAE Schuchert & Le Vene 1929 

Subfamily DRABOVIINAE Havlicek 1950 

DRABOVIA Havlicek 1950 

Drabovia cf. fascicostata Havlicek 

(PL 14, figs. 14-16, 18, 19) 

1950 Drabovia fascicostata Havlicek : 15. 
1950 Drabovia fascicostatata Havlicek : 47. 
1950 Drabovia fascicostata Havlicek : 47. 

Diagnosis. Ventribiconvex Drabovia ornamented by fascicostellae with about 
5 ribs per mm, 5 mm antero-medially of dorsal umbo. 

Description. Semi-elliptical, ventribiconvex, sulcate Drabovia, with orthogonal 
cardinal angles of young shells becoming obtuse in mature growth stages so that the 
hinge line is less than the maximum width of the shell ; carinate pedicle valve 33% 
as deep as long, brachial valve 73% as long as wide and 15% as deep as long (the 
mean for 4 valves) with an evenly concave sulcus and evenly convex sides ; ventral 
interarea relatively short, curved apsacline, dorsal interarea very short anacline, 
delthyrium and notothyrium open ; radial ornamentation fascicostellate, divided 
into sectors by capillae, with counts of 6 and 5 ribs per mm occurring 5 mm antero- 
medially of the umbones of 1 and 4 brachial valves respectively, costellae branching 
internally in valves up to 5 mm long with ia, 2a, 3a and 4a occurring early but with 
3a°)3ala, 3a°)3C and 4a°)4b present (costellae insertion code of Bancroft, 1945). 

Cardinal process with a very small bulbous myophore and relatively long linear 
shaft, becoming thicker in mature valves, continuous with median ridge-like pos- 
terior part of sulcus ; brachiophores short, acutely divergent, bases 17% as long as 
the brachial valve, parallel or convergent with the median ridge, sockets rarely 
defined by fulcral plates ; muscle scar unknown. 

Teeth small supported by slightly divergent dental plates extending forward for 
16% the length of the pedicle valve ; ventral muscle field subcordate with broad 
adductor scar bounded laterally by lobate diductor scars impressed anteriorly for 
about one-quarter the pedicle valve length. 

Figured material. 

length width (mm) 
External and internal moulds of brachial valve 

(BB 35316a, b) 4-0 4-8 

External and internal moulds of brachial valve 

(BB 35317a, b) 4-5 9-4 

External and internal moulds of pedicle valve (BB 35318a, b) 4-5 5*2 

External and internal moulds of pedicle valve (BB 35319a, b) 5-5 7-5 

Horizon and locality. Spy Wood Grit exposed 1100 yds NNE of Rorrington 
(GridRef. S J 303015). 



88 SHELVE DISTRICT 

Discussion. The first Drabovia recorded in the British Isles is represented by a 
small number of moulds (Tables 51-54) which compare closely with D. fascicostata 
Havlicek (1950 : 47) from the Caradocian Letna Beds of Czechoslovakia. The 
ribbing of the Shropshire specimens may prove to be finer and the pedicle valve more 
convex but, provisionally at least, these differences are less important than the many 
points of resemblance. 

Table 51 

Statistics of length (1) and maximum width (w) of 17 brachial valves of Drabovia cf. fascicostata 

Havlicek 

1 mm (var 1) 2-71 (1-192) 

w mm (var w) 3*71 (1-361) 

r 0956 

a (var a) 1-1068 (0-00647) 

Table 52 

Statistics of length (1) and maximum depth (th) of 7 pedicle valves of Drabovia cf. fascicostata 

Havlicek 

1 mm (var 1) 3-21 (1-231) 

th mm (var th) 1 -06 (0-069) 

r 0-988 

a (var a) 0-2375 (0-00028) 

Table 53 

Statistics of length (1) of 8 pedicle valves and the length of dental plates (dl) of Drabovia cf. 

fascicostata Havlicek 

1 mm (var 1) 3-47 (0-931) 

dl mm (var dl) 0-57 (0-043) 

r 0-837 

a (var a) 0-2161 (0-00233) 

Table 54 

Statistics of length (1) and the length of brachiophore bases (lc) in 8 brachial valves of Drabovia 

cf . fascicostata Havlicek. 

lmm (var 1) 3-52 (0-831) 

lc mm (var lc) 0-59 (0-015) 

r 0-820 

a (var a) 0-1367 (0-00102) 

NOCTURNIELLA Havlicek 1950 

Nocturniella sp. 

(PI. 14, figs. 17, 20) 

The internal mould (BB 35341) of a brachial valve, about 4 mm long and 4-5 mm 
wide, is the only draboviinid so far recovered from the Mytton Flags -70 yds north 



ORDOVICIAN BRACHIOPODA 89 

of the entrance to Yewtree Level, The Hollies, Snailbeach (Grid Ref. S J 380018). 
The valve was subcircular in outline and about one-sixth as deep as long with evenly 
convex lateral areas and a strong median sulcus, represented by a broad median 
ridge on the valve interior, becoming fainter anteriorly. Internal impressions of 
ribbing indicate that the valve tended to be multicostellate with 5 costellae per mm 
at the antero-medial margin. The short planar interarea is anacline, divided medially 
by a notothyrium with a low platform bearing a linear cardinal process, and bounded 
by short pointed brachiophores with subparallel bases extending forward for one- 
eighth the length of the valve and sockets defined by fulcral plates. The adductor 
muscle scar impressions are faint and incomplete. 

In the absence of impressions of the pedicle valve, identification of the specimen 
is not a certainty. However, the presence of a sulcus and the subparallel attitude 
of the brachiophore bases favour its assignment to Nocturniella. Yet it is quite 
distinct from the only described species of that genus, N. nocturna Barrande from 
the Arenig Komarov Beds of Czechoslovakia (Havlicek 1950 : 125), especially in the 
plate-like aspect of its cardinal process and the anterior suppression of the sulcus. 



Family DALMANELLIDAE Schuchert 1913 

DALMANELLA Hall & Clarke 1892 

Dalmanella parva Williams emended 

(PL 15, figs. 1-4, 7) 

1949 Dalmanella parva Williams : 169. 

Diagnosis. Small, ventribiconvex Dalmanella with a brachial valve 77% as 
long as wide, and with fine costellae, rarely branching externally, numbering about 
6 per mm at the 2 mm growth stage ; dental plates 23% as long as the pedicle valve 
and 130% as divergent as long ; subparallel brachiophore bases 22% as long as the 
brachial valve and separated from each other by 97% of their length. 

Description. Small, ventribiconvex Dalmanella with obtuse cardinal angles ; 
pedicle valve 32% as deep as long with a carinate rounded median zone and evenly 
sloping lateral areas ; brachial valve 77% as long as wide and about one-fifth as deep 
as long with a shallow median sulcus almost three-quarters as wide as valve length 
flanked by evenly convex lateral areas ; ventral interarea slightly curved apsacline 
longer than anacline dorsal interarea, delthyrium and notothyrium open, pedicle 
callist usually conspicuous ; radial ornamentation costellate commonly 6 per mm, 
2 mm antero-medially of the dorsal umbo ; branching simple, rarely developing 
secondaries especially externally (only 3 external costellae observed in sectors III 
and IV of 24 well-preserved exteriors). 

Teeth small, supported by dental plates extending anteriorly for 23% the length 
of the pedicle valve and diverging for 130% their length ; ventral muscle field 
bilobed with diductor scars extending anteriorly for 35% the length of the pedicle 
valve but not surrounding the median adductor field ; vascula media slightly diver- 
gent, remaining mantle canal system obscure. 



go SHELVE DISTRICT 

Cardinal process consisting of linear shaft and small rounded myophore ; brachio- 
phores short divergent, with subparallel bases flanking well-developed notothyrial 
platform extending anteriorly for 22% of the length of the brachial valve and sep- 
arated from each other by 97% of their length ; adductor scars poorly differentiated, 
suboval, situated on either side of low median ridge and extending anteriorly for 
59% of the length of the brachial valve. 

Figured material. 

length width (mm) 
Internal mould of brachial valve (BB 35434) 4-0 

Internal mould of pedicle valve (BB 35435) 3-5 - 

Internal mould of brachial valve (BB 35436) 3-0 

External mould of brachial valve (BB 35437) 4-0 - 

Horizon and locality. Lower Llandeilo yellow-weathering calcareous sand- 
stones exposed in quarry 300 yds west of Ysgubor-wen Farm, Llandeilo (Grid Ref. 
SN 644224). 



Dalmanella salopiensis sp. nov. 
(PL 15, figs. 5, 6, 8-12) 

Diagnosis. Small, ventribiconvex Dalmanella with a brachial valve JJ°/ as 
long as wide and 15% as deep as long and fine costellae, rarely branching externally, 
commonly numbering 6 per mm at the 2 mm growth stage ; dental plates 21% as 
long as the pedicle valve and 159% as divergent as long ; ventral muscle scar extend- 
ing anteriorly for 35% of the length of the pedicle valve ; subparallel brachiophore 
bases 22% as long as the brachial valve and separated from each other by 86% of 
their length. 

Description. Small, subquadrate, ventribiconvex Dalmanella with obtuse 
cardinal angles ; pedicle valve 38% as deep as long with rounded carina medially 
and evenly sloping lateral areas, brachial valve 77% as long as wide and 15% as 
deep as long with shallow median sulcus about four-fifths as wide as valve length 
flanked by evenly convex lateral areas ; ventral interarea slightly curved apsacline, 
longer than anacline dorsal interarea, delthyrium and notothyrium open ; radial 
ornamentation of costellae commonly 6 per mm 2 mm antero-medially of the dorsal 
umbo ; branching simple, rarely developing externally with only 4 external ribs 
observed in 32 mature brachial valves. 

Teeth supported by dental plates extending anteriorly for 21% the length of the 
pedicle valve and extending laterally to 159% of their length ; ventral muscle scar 
bilobed extending anteriorly for 35% of the length of the pedicle valve ; mantle 
canal system unknown. 

Cardinal process consisting of linear shaft and compact myophore on a well- 
developed notothyrial platform ; brachiophores short, divergent with subparallel 
bases extending anteriorly for 22% of the length of the brachial valve and separated 



ORDOVICIAN BRACHIOPODA 



9i 



from each other by 86% of their length ; adductor scars suboval, faintly impressed 
on either side of low median ridge and extending anteriorly for 59% of the length of 
the brachial valve. 

Type material. 

Holotype External and internal moulds of brachial valve 

(BB 35446a, b) 
Paratypes External and internal moulds of pedicle valve 

(BB 35447a, b) 

External and internal moulds of brachial valve 

(BB 35448a, b) 

External and internal moulds of brachial valve 

(BB 35449^, b) 

External and internal moulds of brachial valve 

(BB 3545oa, b) 

Type horizon and localities. Meadowtown Beds : BB 35446 from loose flags 
by the side of Minicop Farm (Grid Ref . SJ 314018) ; BB 35450 from flags in Ouinton's 
Quarry in the field 200 yds north-east of the chapel at Meadowtown (Grid Ref. 
SJ 312013) ; BB 35448-9 from tuffaceous flags 100 yds south-east of Minicop 
Farm (Grid Ref. SJ 315018) ; BB 35447 from shales along Meadowtown- Rorrington 
road 220 yds from Meadowtown Chapel (Grid Ref. SJ 309012). 



:ngth 


width (mm) 


3-0 


3-8 


3-2 


37 


i-6 


2-4 


17 


2-5 


2-5 


3-2 



Dalmanella salopiensis gregaria sp. et subsp. nov. 
(PI. 15, figs. 13-16, 20) 

Diagnosis. Like Dalmanella salopiensis but with a brachial valve 74% as long as 
wide and 16% as deep as long ; dental plates 19% as long as the pedicle valve and 
184% in lateral extent as long ; ventral muscle scar 36% as long as the pedicle 
valve ; subparallel brachiophore bases 23 % as long as the brachial valve and separated 
from each other by 91% of their length. 

Type material. 

Holotype External and internal moulds of brachial valve 

(BB 35443a, b) 
Paratypes External and internal moulds of pedicle valve 

(BB 35444a, b) 

External and internal moulds of brachial valve 

(BB 35445a, b) 

Internal and external moulds of pedicle valve 

(BB 35407a, b) 



length width (mm) 
2-5 3-5 



2-4 



3-8 



2-8 

2-8 

5-o 



92 SHELVE DISTRICT 

Type horizon and locality. Spy Wood Grit exposed in north bank of Spy Wood 
Brook, 170 yds north-east of Spy Wood Cottage (Grid Ref. SO 282958). 



Dalmanella salopiensis transversa sp. et subsp. nov. 

(Pi. 15, figs. 17-19, 21-24 ; Pi- 16, fig. 1) 

Diagnosis. Like Dalmanella salopiensis but with a brachial valve 73% as long 
as wide and about one-tenth as deep as long ; dental plates 20% as long as the pedicle 
valve and 192% in lateral extent as long ; subparallel brachiophore bases 22% as long 
as the brachial valve. 

Type material. 

length width (mm) 

Holotype External and internal moulds of brachial valve 

(BB 35438a, b) 3-4 4-2 

Paratypes External and internal moulds of pedicle valve 

(BB 35439a, b) 3-5 

External and internal moulds of brachial valve 

(BB 35440a, b) 2-5 3-5 

External and internal moulds of pedicle valve 

(BB 3544ia, b) 2-5 3-0 

External and internal moulds of pedicle valve 

(BB 35442a, b) 2-5 3-2 

Type horizon and locality. Aldress Shales exposed in the bank of Ox Wood 
Dingle at the south-west corner of Ox Wood, just north of the Rorrington-Wotherton 
road (Grid Ref. SJ 290007). 

Discussion. Dalmanella is a common brachiopod in the Shelve successions. 
Apart from a Dalmanella-like species from the Mytton Flags which has been sep- 
arately described, the genus is represented by a series of closely related stocks with 
small adult shells usually between 2 and 5 mm long ornamented by simply arranged 
costellae numbering 6 per mm 2 mm antero-medially of the dorsal umbo, and bearing 
a cardinal process with a small rounded myophore, well-developed subparallel 
brachiophore bases and the impressions of a bilobed ventral muscle field. These 
stocks occur in the Betton Beds to Aldress Shales inclusive but are very common in 
the Meadowtown Beds, the Spy Wood Grit and the Aldress Shales which have 
provided the data for the three samples listed in Tables 55 to 63. 

At first sight, the stocks appear to be like the Lower Llandeilo Dalmanella parva 
Williams (1949 : 169) in the dimensions and proportions of their external and internal 
features. However, a sample of that species from the type locality proved to be 
significantly different from Shelve Dalmanella in a number of attributes. From all 
three samples D. parva differs in the rate of anterior expansion and in the inherently 
more acute divergence of its dental plates. The species further differs from the 
Meadowtown and Spy Wood forms in the relatively faster growth in width and, from 



ORDOVICIAN BRACHIOPODA 93 

the Aldress sample, in the inherently greater width of the brachial valve. The 
Llandeilo species also differs from the Meadowtown and Aldress stocks in the shape 
parameters and relative growth of the cardinalia. 

In contrast to these decisive specific differences, the relationship between the three 
Shelve samples is much more complicated. The Meadowtown sample, herein 
recognized as the new species D. salopiensis, differs from both the Spy Wood and 
Aldress stocks in the more acutely divergent attitude of the dental plates ; and 
additionally from the former in the residual length of the ventral muscle scar 
and the divergence of the brachiophore bases. It differs from the latter in the 
relative width of the brachial valve. Moreover there is also a significant difference 
(p < o-ooi) in the residual shape of the brachial valve of the Aldress and Spy Wood 
samples. These relationships are most conveniently expressed by assuming that 
the Aldress and Spy Wood samples constitute two new subspecies, D. salopiensis 
transversa and D. salopiensis gregaria respectively. 

The stratigraphic ranges of these taxa are well defined. The small number of 
moulds collected from the Betton Beds are indistinguishable from D. salopiensis. 
A larger sample of specimens from the Rorrington Beds, on the other hand, differs 
from D. salopiensis at least in the degree of divergence of the dental plates (p < o-ooi), 
but they proved to be identical in every respect with the new Spy Wood subspecies 
D. salopiensis gregaria. 

Neither D. salopiensis (s.l.) nor D. parva are like other Anglo- Welsh Dalmanella. 
D. indica (Whittington) appears to be most closely related, but comparisons of 
statistical data for that species (Williams 1963 : 382-385) with those for D. parva 
and D. salopiensis (s.l.) reveals significant differences in a number of features in- 
cluding the parallel disposition of the brachiophore bases in the older species and 
especially in the simplicity of their ribbing patterns which are essentially : ia, 1, 
2a, 2, 3a, 3b, 3, 4al, 4a, 4, 4a . 

Table 55 

Statistics of length (1) and maximum width (w) of n brachial valves of Dalmanella parva Williams 
(A), D. salopiensis sp. nov. (B), D. salopiensis gregaria sp. et subsp. nov. (C) and D. salopiensis 

transversa sp. et subsp. nov. (D) 





A 


B 


C 


D 


n 


42 


82 


46 


14 


1 mm 


2-76 


2-06 


2-28 


2-62 


(var 1) 


(o-54) 


(0-461) 


(o-443} 


(0-814) 


w mm 


3-57 


2-73 


3-o9 


3-59 


(var w) 


(o795) 


(0-522) 


(o-55) 


(1-148) 


r 


0986 


0-972 


o-955 


o-993 


logel 


0-9808 


0-6721 


0-7845 


0-9078 


(var logel) 


(0-0685) 


(0-1029) 


(0-0815) 


(0-1119) 


log e W 


1-2429 


0-9708 


1 -099 


1-2364 


(var log e w) 


(0-0603) 


(0-0675) 


(0-0561) 


(0-0851) 


r e 


0-989 


0-976 


o-959 


o-993 


a 


0-9384 


o-8i 


0-8294 


0-8723 


(var a) 


(0-00049) 


(0-00038) 


(0-00126) 


(0-00088 



94 



SHELVE DISTRICT 



Table 56 

Statistics of length (1) and thickness (th) of n pedicle valves of Dalmanella parva Williams (A), 
D. salopiensis sp. nov. (B), D. salopiensis gregaria sp. et subsp. nov. (C) and D. salopiensis 

transversa sp. et subsp. nov. (D) 





A 


B 


C 


D 


n 


29 


72 


53 


8 


1 mm 


3-29 


2-37 


3-i8 


2-82 


(var 1) 


(o-86 3 ) 


(0-608) 


(1-223) 


(0-476) 


th mm 


I -04 


0-91 


i-35 


I-OI 


(var th) 


(0-092) 


(0-089) 


(0-0182) 


(0-038) 


r 


0-909 


o-88i 


o-933 


0-885 


a 


0-3274 


0-3827 


0-3861 


0-2839 


(var a) 


(0-00069) 


(0-00046) 


(0-00037) 


(0-00291) 



Table 57 

Statistics of length (1) of n pedicle valves and length of dental plates (dl) of Dalmanella parva 

Williams (A), D. salopiensis sp. nov. (B), D. salopiensis gregaria sp. et subsp. nov. (C) and D. 

salopiensis transversa sp. et. subsp. nov. (D) 





A 


B 


c 


D 


n 


52 


102 


69 


11 


1 mm 


3-35 


2-22 


2-92 


2-69 


(var 1) 


(0-654) 


(0-609) 


(1-159) 


(0-685) 


dl mm 


0-78 


0-46 


o-57 


o-53 


(var dl) 


(0-061) 


(0-025) 


(0-045) 


(0-026) 


r 


0-882 


o-88 


0-907 


0-919 


logel 


1-1807 


0-7380 


1-0063 


0-9446 


(var logel) 


(0-0564) 


(0-1168) 


(0-1279) 


(0-0902) 


logedl 


— 0-2978 


-0-8268 


— 0-6292 


-0-6749 


(var logedl) 


(0-0957) 


(0-1104) 


(0-13) 


(0-0868) 


r e 


o-888 


0-887 


0-913 


0-925 


a 


1-3026 


0-9724 


1-0082 


0-9808 


(var a) 


(0-00717) 


(0-0020I) 


(0-00253) 


(0-01538) 



ORDOVICIAN BRACHIOPODA 95 



Table 58 

Statistics of length (1) and maximum anterior divergence (w) of the dental plates in n pedicle 

valves of Dalmanella parva Williams (A), D. salopiensis sp. nov. (B), D. salopiensis gregaria 

sp. et subsp. nov. (C) and D. salopiensis transversa sp. et subsp. nov. (D) 





A 


B 


C 


D 


n 


53 


77 


56 


8 


1 mm 


079 


0-46 


o-55 


o-53 


(var 1) 


(0-07) 


(0-024) 


(0-046) 


(0-02I) 


w mm 


1-03 


o-73 


I-OI 


1-02 


(var w) 


(0-069) 


(0-042) 


(0-087) 


(0-059) 


r 


0-951 


0-805 


0-879 


O-gol 


logel 


— 0-2842 


-0-8385 


-0-6618 


-0-6689 


(var logel) 


(0-1054) 


(0-1088) 


(o-Mi) 


(0-0729) 


log e W 


-0-0055 


-0-3462 


— 0-0282 


— 0-0028 


(var log e w) 


(0-0631) 


(0-075) 


(0-0812) 


(0-0548) 


r e 


o-955 


0-813 


o-886 


0-907 


a 


0-774 


0-8306 


0-7588 


0-8669 


(var a) 


(0-00103) 


(0-00312) 


(0-00229) 


(0-022I7) 



Table 59 

Statistics of length (1) of n pedicle valves and length of ventral muscle scar (sc) of Dalmanella 

parva Williams (A), D. salopiensis sp. nov. (B), D. salopiensis gregaria sp. et subsp. nov. (C) and 

D. salopiensis transversa sp. et subsp. nov. (D) 





A 


B 


C 


D 


n 


44 


69 


61 


8 


1 mm 


333 


2-44 


3-06 


2-74 


(var 1) 


(o-6o8) 


(o-474) 


(1-099) 


(0-608) 


s"c mm 


1-17 


0-83 


I-I2 


0-87 


(var sc) 


(0-084) 


(0-061) 


(o-I 9 8) 


(0-088) 


r 


0-914 


0-853 


0-899 


0-969 


log e l 


I-I775 


0-8558 


I-o6l9 


0-968 


(var logel) 


(0-0532) 


(0-0763) 


(0-III2) 


(0-078) 


logeSC 


0-1317 


-0-2313 


0-0388 


— 0-1877 


(var logeSc) 


(0-0592) 


(0-0858) 


(0-1468) 


(0-1086) 


r e 


0-92 


o-86i 


0-905 


0-9713 


a 


1-0554 


1-06 


1-1487 


1-1804 


(var a) 


(0-00408) 


(0-00434) 


(0-00403) 


(0-01315) 



96 shelve district 

Table 60 

Statistics of length (1) of n brachial valves and length of brachiophore bases (lc) of Dalmanella 

parva Williams (A), D. salopiensis sp. nov. (B), D. salopiensis gregaria sp. et subsp. nov. (C) and 

D. salopiensis transversa sp. et subsp. nov. (D) 





A 


B 


C 


D 


n 


38 


121 


55 


22 


1 mm 


3" 


2-18 


2-56 


2-48 


(var 1) 


(o-473) 


(0-517) 


(o-474) 


(o-8n) 


lc mm 


0-69 


0-49 


0-58 


o-55 


(var lc) 


(0-027) 


(0-025) 


(0-023) 


(0-027) 


r 


0-814 


091 1 


0-884 


0-917 


log e l 


i-iii8 


0-726 


0-906 


0-8472 


(var logel) 


(0-0476) 
-0-3913 


(0-1036) 
-0-7675 


(0-0695) 
-0-5684 


(0-1236) 


logelc 


— 0-6404 


(var logelc) 


(0-0543) 


(0-1005) 


(0-0661) 


(0-0851) 


r e 


0-822 


0-915 


0-889 


0-922 


a 


1-068 


0-985 


0-9751 


0-8299 


(var a) 


(0-01028) 


(0-00132) 
Table 61 


(0-00376) 


(0-00513) 



Statistics of length (1) and maximum lateral extension (w) of the brachiophore bases in n brachial 

valves of Dalmanella parva Williams (A), D. salopiensis sp. nov. (B), D. salopiensis gregaria sp. et 

subsp. nov. (C) and D. salopiensis transversa sp. et subsp. nov. (D) 





A 


B 


C 


D 


n 


46 


118 


52 


21 


1 mm 


0-65 


0-49 


0-58 


o-57 


(var 1) 


(0-025) 


(0-025) 


(0-023) 


(0-022) 


w mm 


0-63 


0-42 


o-53 


0-48 


(var w) 


(0-02I) 


(0-008) 


(0-013) 


(0-013) 


r 


0-778 


0-771 


0-596 


0-741 


logel 


— 0-4617 


-0-765 


-0-5718 


— 0-6008 


(var logel) 


(0-0587) 


(0-0993) 


(0-0667) 


(0-0656) 


log e W 


— 0-4811 


-0-8848 


-0-665 


-0-7653 


(var log e w) 


(0-0496) 


(0-0422) 


(0-0452) 


(0-0567) 


r e 


0-787 


0-781 


0-604 


o-75 


a 


0-9I88 


0-652 


0-8228 


09303 


(var a) 


(0-0073) 


(0-00142) 

Table 62 


(0-00859) 


(0-01993) 



Statistics of length (1) of n brachial valves and length of adductor scars from umbo (sc) of 
Dalmanella parva Williams (A), D. salopiensis sp. nov. (B) and D. salopiensis gregaria sp. et 

subsp. nov. (C) 

ABC 
n 13 24 22 

1mm (var 1) 3-45(0-411) 2-74(0-505) 2-74(0-5) 

s"c mm (var sc) 1-99 (0-207) 1-56 (0-149) i-6 (0-152) 

r 0-906 0-925 0929 

a (var a) 0-7104 (0-00821) 0-5439 (0-00194) °'55 I 8 (0-00207) 



ordovician brachiopoda 97 

Table 63 

The distribution of brachial valves of Dalmanella parva Williams (A), D. salopiensis sp. nov. (B), 

D. salopiensis gregaria sp. et subsp. nov. (C) and D. salopiensis transversa sp. et subsp. nov. (D) 

with 4-8 costellae per mm, 2 mm antero-medially of the dorsal umbones 

Costellae per mm 
45678 
A 14863 

B 1 11 17 3 - 

C 2 5 9 

D -222- 



Dalmanella elementaria sp. nov. 
(PL 16, figs. 2-8) 

Diagnosis. Very small, ventribiconvex, sulcate Dalmanella with a finely costel- 
late ornamentation, faintly impressed bilobed ventral muscle field, slightly divergent 
dental plates, a short ridge-like cardinal process, subtriangular blade-like brachio- 
phores with subparallel bases and small fulcral plates. 

Description. Very small, ventribiconvex, subquadrate Dalmanella with rounded, 
obtuse cardinal angles ; pedicle valve slightly carinate medially, one-third as deep 
as long ; gently convex brachial valve very rarely more than 3 mm long with a mean 
length relative to width of 78-3% (variance 109-2) for 5 valves, median sulcus usually 
persistent, rounded, almost three-fifths as wide as the valve length, bounded by 
evenly convex lateral areas ; ventral and dorsal interareas short, apsacline and 
anacline respectively, with an open delthyrium containing a small pedicle callist 
and an open notothyrium ; ribbing very fine with 7 and 8 costellae per mm, 2 mm 
antero-medially of the umbones of 5 and 2 brachial valves respectively, first three 
sectors narrow with simple branching normally consisting of ia, 1, 2a, 2b, 2, 3a!, 3a, 

3, 3a°. 

Teeth supported by slightly divergent dental plates extending forward for an 
average of 26-3% (variance 46-8) the length of 5 valves ; ventral muscle field rarely 
and faintly impressed, apparently bilobed, with shorter median adductor scar 
flanked by lobate diductors. 

Cardinal process consisting of a low impersistent ridge not extending to posterior 
margin of notothyrium, brachiophores subtriangular, blade-like with subparallel 
bases extending anteriorly for one-fifth the length of the valve (mean lengths and 
variances of 13 valves and brachiophore bases = 1-99 (0-256) and 0-41 (0-019) 
respectively ; r = 0-943) ; sockets narrow, defined by small fulcral plates. 

Type material. 

length width (mm) 
Holotype External and internal moulds of brachial valve 

(BB 35352a, b) 3-2 3-8 

7 



98 SHELVE DISTRICT 

length width (mm) 
Paratypes External and internal moulds of pedicle valve 

(BB 35353a, b) 2-5 3-5 

External and internal moulds of pedicle valve 

(BB 35354a, b) i-8 

External and internal moulds of brachial valve 

(BB 35355a, b) 3-0 

External and internal moulds of brachial valve 

(BB 35356a, b) 2-0 3-0 

External and internal moulds of brachial valve 

(BB 35357a, b) 2-5 

Type horizon and localities. Mytton Flags (Shelve Church Member) : 
BB 35353a, b from exposures in road level north of Wood House (Grid Ref . S J 338003) , 
all other specimens from a trench cut 1 ft above road level near Shelve Church (Grid 
Ref. SO 335990). 

Discussion. Among the fossils recovered from the Mytton Flags, there quite 
commonly occur impressions of a small costellate, thin-shelled orthide which have 
all the essential characters expected of early enteletaceans. The combination of 
such features is so striking that, although there is, as yet, no proof, such as the pre- 
sence of microscopic moulds of puncta, of their enteletacean affinities, they have been 
identified as primitive Dalmanella. Certainly they are strongly reminiscent of 
young Dalmanella in many respects, yet there are characters that immediately 
distinguish these Arenig forms from all other described Dalmanella. Among these 
are the fineness and simple arrangement of costellae and the ridge-like cardinal 
process which did not extend the length of the notothyrial platform nor become 
differentiated into myophore and shaft as in other species assigned to the genus. 
Not only do these features render the new species unique among known Dalmanella 
but, together with the obscurity of muscle impressions and simplicity of brachio- 
phore arrangement, are also the features one would expect to find in a stock that was 
ancestral to the more typical Ordovician Dalmanella. Indeed when more is known 
about these Arenigian stocks, they may prove to be so distinctive in the generalized 
nature of their cardinalia as to constitute a distinct species group within the 
Dalmanella taxon. 



ONNIELLA Bancroft 1928 

Onniella ostentata Williams lepida subsp. nov. 

(PI. 16, figs. 9-14) 

Diagnosis. Like Onniella ostentata Williams (1963 : 405) but with finer costellae 
numbering 4 to 6 per mm at the 5 mm growth stage, shorter dental plates which 
grew forward more slowly and a massive bilobed cardinal process not united to the 
brachiophores in adult shells. 



ORDOVICIAN BRACHIOPODA 99 

Description. Small, ventribiconvex Onniella with obtusely rounded cardinal 
angles ; pedicle valve 23% as deep as long with a subcarinate rounded median zone 
and evenly sloping to slightly concave lateral areas, brachial valve 76% as wide and 
11% as deep as long with a shallow narrow to weak sulcus ; ventral interarea planar 
apsacline, longer than anacline dorsal interarea, delthyrium and notothyrium open, 
pedicle callist usually well developed ; radial ornamentation of costellae commonly 
4 per mm, 5 mm antero-medially of dorsal umbo, external costellae poorly developed 
in sectors III and IV. 

Teeth small trigonal, supported by dental plates extending anteriorly for 16% of 
the length of the pedicle valve and diverging for 49% of their length ; ventral muscle 
field bilobed with the diductor scars extending anteriorly for 29% of the length of 
the pedicle valve and flanking a shorter submedial adductor scar ; vascula media 
proximally divergent, remaining mantle canal system unknown. 

Cardinal process consisting of linear shaft and rounded myophore in early growth 
stages but usually massive and medially cleft in valves more than 3 mm long ; 
brachiophores short, acutely divergent, with bases, which may be broad and in- 
dented by a pit, extending forward for 19% of the length of the brachial valve and 
splaying laterally for 58% of their length ; sockets well defined but without fulcral 
plates ; adductor scars suboval, impressed on either side of median ridge and 
extending anteriorly for 57% of the length of the valve ; mantle canal system 
unknown. 

Type material. 

length width (mm) 
Holotype External and internal moulds of brachial valve 

(BB 35456a, b) 4-3 5-0 

Paratypes External and internal moulds of pedicle valve 

(BB 35455a, b) 4-8 6-o 

External and internal moulds of brachial valve 
(BB 35457a, b) 3-0 4-0 

Internal mould of brachial valve (BB 35458) 4-0 

Type horizon and locality. Aldress Shales exposed in the bank of Ox Wood 
Dingle at the south-west corner of Ox Wood, just north of the Rorrington-Wotherton 
road (Grid Ref. SJ 290007). 

Discussion. Except for the complementary moulds of a brachial valve 
(BB 35572a, b) collected from the Spy Wood Grit (PI. 16, figs. 15, 18), Onniella is 
restricted in the Shelve area to the Aldress and Whittery Shales. Although it is not 
very common in either formation, two small samples were available for study. They 
proved to be alike in every respect (Tables 64-72) and obviously represent the same 
species. They are however significantly different from penecontemporaneous 
Onniella, such as 0. soudleyensis (Bancroft 1945 : 210) and 0. ostentata (Williams 
1963 : 405). In comparison with the former species, the Shelve Onniella differs in 
the more elongate growth of the brachial valve, the relative length and narrowness 
of the cardinalia and in the early development of a massive bilobed cardinal pro- 
cess. In respect of the last-named feature the Shelve Onniella is nearer 0. ostentata 



ioo SHELVE DISTRICT 

s.s. except that the cardinal process always remains discrete from the brachiophores. 
In other features, too, there is a noteworthy likeness between the two stocks which 
differ only in the relative shortness of the dental plates and the fineness of ribbing 
and poor development of secondary externals in sectors III and IV in the Shelve 
samples. In the Whittery sample, 4, 5 and 6 costellae per mm were counted 5 mm 
antero-medialry of the umbones of 3, 1 and 1 brachial valves ; while 3ala)3a°, 3c)3a° 
and 4b)4b° occurred in 0/4, 1/5 and 1/2 brachial valves respectively. These 
differences seem to be important enough to merit subspecific recognition. 

Table 64 

Statistics of length (1) and maximum width (w) of n brachial valves of Onniella ostentala lepida 
subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 

A B 

n 11 22 

ljnm (var 1) 3-2 (0-638) 3-7 (1-149) 

w mm (var w) 4-12 (0-748) 4-99 (2-006) 

r 0-985 0-962 

log e l (var log e l) I-I 33 I (0-0604) 1-2694 (0-0804) 

logew (var logew) 1-394 (0-0429) 1-5679 (0-0775) 

r e 0-99 0-965 

a (var a) 0-843 (0-00158) 0-9815 (0-00329) 

Table 65 

Statistics of length (1) and depth (th) of 12 brachial valves of Onniella ostentata lepida subsp. nov. 

from the Whittery Shales 

1 mm (var 1) 4-22 (0-532) 

th mm (var th) 0-48 (0-012) 

r 0-865 

a (var a) 0-1527 (0-00058) 

Table 66 

Statistics of length (1) and depth (th) of 10 pedicle valves of Onniella ostentata lepida subsp. nov. 

from the Whittery Shales 

1 mm (var 1) 3-91 (0-761) 

th mm (var th) o-88 (0-068) 

r 0-493 

a (var a) 0-3 (0-00851) 

Table 67 

Statistics of length (1) of n pedicle valves and length of dental plates (dl) of Onniella ostentata 
lepida subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 





A 




B 


n 

1 mm (var 1) 

dl mm (var dl) 

r 
a (var a) 


9 

3-42 (1-017) 
o-57 (0-034) 
0-799 
0-184 (o-ooi 


75) 


12 

3-91 (1-788) 
o-6 (0-061) 
0-909 
0-185 (0-00059) 



ordovician brachiopoda 101 

Table 68 

Statistics of length (1) and maximum anterior divergence (w) of the dental plates in n pedicle valves 
of Onniella ostentata lepida subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 

A B 

n 99 

1 mm (var 1) 0-57 (0-034) °*5 2 (0-026) 

w mm (var w) 1-21 (0-081) i-oi (0-079) 

r 0-858 0-877 

a (var a) I-534& (0-0885) 17305 (0-09891) 



Table 69 

Statistics of length (1) of n pedicle valves and length of ventral muscle scar of Onniella ostentata 
lepida subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 

A B 

n 9 10 

1 mm mm (var 1) 3-42 (1-017) 3'99 ( I- 774) 

sc mm (var sc) 1-04 (0-109) i"U (0-225) 

r 0-915 0-882 

a (var a) 0-3268 (0-00248) 0-3564 (0-00352) 



Table 70 

Statistics of length (1) of n brachial valves and length of brachiophore bases (lc) of Onniella 

ostentata lepida subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 

A B 

n 16 21 

lmm (var 1) 3-19 (0-652) 4-13 (0-981) 

lc mm (var lc) o-66 (0-021) 0-69 (0-027) 

r 0681 0-913 

a (var a) 0-1812 (0-00125) 0-1655 (0-00023) 



Table 71 

Statistics of length (1) and maximum lateral extension (w) of the brachiophore bases in n brachial 
valves of Onniella ostentata lepida subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 

A B 

n 16 15 

1 mm (var 1) o-66 (0-021) o-66 (0-025) 

w mm (var w) 1-1(0-075) 1-19(0-146) 

r 0-803 0-967 

log e l (var log e l) —0-4405(0-048) —0-4437(0-0566) 
log e w (var log e w) 0-0654 (°"°597) 0-1279 (0-0977) 

r e o-8n 0-972 

a (var a) 1-1157(0-03045) 1-3139(0-00719) 



102 shelve district 

Table 72 

Statistics of length (1) of n brachial valves and length of adductor scars (sc) of Onniella ostentata 
lepida subsp. nov. from the Aldress Shales (A) and Whittery Shales (B) 

A B 

n 15 10 

1 mm (var 1) 3-2 (0-696) 4-27 (1-438) 

s"c mm (var sc) 1-88 (0-235) 2-3 (0-4) 

r 0-833 o-979 

a (var a) 0-581 (0-00794) 0-5274 (0-00146) 

Family HARKNESSELLIDAE Bancroft 1928 
HARKNESSELLA Reed 1917 
Harknessella cf. subplicata Bancroft 
(PL 16, fig. 16) 

An incomplete internal mould (BB 35427), about 3 mm long, from the Spy Wood 
Grit exposed on top of the ridge 1440 yds NNE of Rorrington bench mark 599 
(Grid Ref. SJ 303018), appears to represent the immature brachial valve of a 
Harknessella. The semi-oval valve was about three-fifths as long as wide and almost 
one-fifth as deep as long with a well-developed median sulcus over half as wide as the 
valve length and gently convex lateral areas. Ribbing impressions on the mould 
reach almost to the cardinalia and suggest a coarsely costellate arrangement with 
about 4 per mm towards the anterior margin. The cardinalia consisted of a cardinal 
process differentiated into a stout shaft and crenulated myophore, obliquely disposed 
sockets defined by fulcral plates and brachiophores supported by acutely divergent 
bases indented submedially by deep adductor pits. 

The development of adductor pits and the disposition of the brachiophore bases 
and fulcral plates to give the 'angulated subparallel pre-socket lines' of Bancroft 
(1945 : 225) indicate that the brachial valve belonged to Harknessella ; the relative 
coarseness of the ribbing further suggests its close affinities with the species H. 
subplicata Bancroft (1945 : 226) from the Coston Sandstone of E. Shropshire. 

HORDERLEYELLA Bancroft 1928 

Horderleyella cf. plicata Bancroft 

(PL 16, figs. 17, 19, 20 ; PL 17, fig. 1) 

Diagnosis. Small, subquadrate Horderleyella ornamented by 3 to 5 fasci- 
costellae per mm, 2 mm antero-medially of the dorsal umbo ; bilobed ventral 
muscle scar extending anteriorly for about two-fifths the valve length ; cardinal 
process blade-like, subparallel brachiophores extending forward for about 12% of 
the valve length and dorsal adductor scars impressed posteriorly in shallow hollows 
in the notothyrial platform. 

Description. Small, subquadrate, ventribiconvex Horderleyella with cardinal 
angles approximating to 90 ° ; pedicle valve carinate medially, 30% as deep as long ; 



ORDOVICIAN BRACHIOPODA 103 

brachial valve 76% as long as wide and 12% as deep as long with a strong median 
sulcus about two-thirds as wide as the valve length, flanked by flattened lateral 
areas ; ventral and dorsal interareas planar, apsacline and anacline respectively with 
open delthyrium and notothyrium ; radial ornamentation fascicostellate with sub- 
angular ribs branching internally and numbering 3, 4 and 5 per mm, 2 mm antero- 
medially of the umbones of 1, 2 and 1 brachial valves. 

Teeth supported by divergent dental plates extending forward for 22% of the 
length of the valve ; ventral muscle scar bilobed, about as wide as long, with sub- 
triangular diductor impressions extending forward for 40% the length of the valve 
but not enclosing the shorter median adductor scar. 

Cardinal process simple, blade-like, brachiophores narrowly divergent with sub- 
parallel bases extending forward for 18% of the length of the brachial valve and 
deeply separated from well-developed fulcral plates defining splayed sockets ; 
adductor scars lightly impressed on either side of the wide median ridge except 
posteriorly where they are inserted as a pair of shallow hollows into the notothyrial 
platform. 

Figured material. 

length width (mm) 
External and internal moulds of pedicle valve (BB 35431a, b) 6«o 6*5 

Internal mould of brachial valve (BB 35432) 3-5 4-0 

Internal mould of brachial valve (BB 35433) 4-0 

Horizons and localities. Spy Wood Grit : BB 35431 from exposures on top 
of ridge 1440 yds NNE of Rorrington bench mark 599 (Grid Ref. SJ 303018) : 
BB 35432 from exposures 1100 yds north-east of Rorrington (Grid Ref. SJ 303015). 
Aldress Shales : BB 35433 from exposures in bank of Ox Wood Dingle at south-west 
corner of Ox Wood a few yards north of the Rorrington-Wotherton road (Grid Ref. 
SJ 290007). 

Discussion. A small number of moulds of Horderleyella, mainly those of brachial 
valves, have been recovered from the Spy Wood Grit and Aldress Shales. They 
appear to be conspecific with H. plicata Bancroft (1928 : 186 ; 1945 : 236) from the 
Costonian and Harnagian successions of E. Shropshire, especially in the radial orna- 
mentation and the shape and internal morphology of the brachial valve. The differ- 
entiation of the cardinal process of H. plicata into myophore and shaft is probably 
not a diagnostic feature in view of the much smaller size of the Shelve specimens. 
Such differences in shell size may also account for the relative delicacy of the car- 
dinalia of H. subcarinata MacGregor (1961 : 193) from the Upper Llandeilo of the 
Berwyn Hills which may also be conspecific with H. plicata. 

In contrast to the close similarity of the brachial valves, the Shelve pedicle valves 
may prove to be relatively deeper and the ventral muscle scar wider and shorter, 
being only 79% as wide as long and 31% of the length of three pedicle valves of H. 
plicata. Larger collections, however, might indicate that these differences fall within 
the range of variability of H. plicata and, provisionally at least, the Shelve material 
can be identified as that species. 



io 4 SHELVE DISTRICT 

Horderleyella sp. 

(PI. 17, figs. 2, 3) 

The external and internal moulds of an immature pedicle valve of Horderleyella, 
2-8 mm long and 3-5 mm wide (BB 35420a, b), have been recovered from the 
Meadowtown Beds exposed in the lane side 370 yds north-west of Meadowtown 
Chapel (Grid Ref. SJ 311015). 

The valve was subquadrate in outline with obtuse cardinal angles and about one- 
quarter as deep as long in the vicinity of a strong narrow median carina although the 
lateral areas are only very gently convex. The radial ornamentation was strongly 
fascicostellate with a count of 4 ribs per mm, 2 mm anterior of the umbo. Inter- 
nally, divergent dental plates, extending anteriorly for almost one-fifth the length of 
the valve, formed the posterior boundaries of a subpentagonal ventral muscle scar 
which was about three-quarters as long as wide and extended forward for about 
three-tenths the valve length. 

The valve cannot be placed unequivocally in any of the described species of Horder- 
leyella. The relatively fine fascicostellate ornamentation is more diagnostic of the 
Upper Llanvirn Horderleyella convexa Williams (1949 : 171), for which counts of 3, 
4 and 5 costellae per mm have been obtained 2 mm anterior of the umbones of 1, 2 
and 2 brachial valves respectively. The muscle scar, however, is very much wider 
than even that of H . convexa ; and, although this difference is likely to be attributable 
to the immaturity of the shell, more material will have to be obtained before specific 
identification is merited. 



REUSCHELLA Bancroft 1928 

Reuschella horderleyensis Bancroft carinata subsp. nov. 

(PL 17, figs. 4-9) 

Diagnosis. Subquadrate Reuschella with a narrowly carinate and laterally 
flattened pedicle valve about 23% as deep as long and a deeply sulcate brachial 
valve 73% as long as wide and 15% as deep as long ; radial ornamentation coarsely 
fascicostellate with 2 or 3 ribs per mm, 10 mm anteriorly of the dorsal umbo and 
poorly developed external branching except in the lateral sectors ; ventral muscle 
scar slightly bilobed and about one-third as long as the pedicle valve ; brachiophore 
bases extending forward for 14% of the length of the brachial valve. 

Description. Ventribiconvex, subquadrate Reuschella with acute cardinal 
angles in specimens up to 14 mm long but usually becoming obtuse in mature shells, 
brachial valve 73% as long as wide and 15% as deep as long with a narrowly rounded 
sulcus having a mean width relative to the length of 4 valves of 47% (range 39% to 
55%) '• pedicle valve with a mean depth relative to length of 23% for 3 specimens 
representing the elevation of a narrow rounded carina flanked by flattened, rarely 
undulate lateral areas ; ventral and dorsal interareas well developed, apsacline and 
anacline respectively with open delthyrium and notothyrium ; radial ornamentation 
coarsely fascicostellate with counts of 2 and 3 ribs per mm, 10 mm anteriorly of the 



ORDOVICIAN BRACHIOPODA 105 

umbones of 4 and 3 brachial valves, external branching poorly developed but pro- 
gressively earlier in lateral sectors. 

Teeth strong with well-developed crural fossettes supported by divergent dental 
plates with an average forward extension relative to the length of 6 pedicle valves 
of 19% (range 17% to 22%) ; ventral muscle scar bilobed, 83% as wide as long in 12 
pedicle valves (range 72% to 104%), extending anteriorly for 34% of the length of 
7 valves (range 28% to 39%) and consisting of a broad median adductor impression 
not enclosed by subtriangular diductor tracks ; mantle canal system probably 
lemniscate. 

Cardinal process consisting of a massive shaft with serrated ridge-like myophore 
embedded in a thickened notothyrial platform indented anteriorly by shallow pits 
for the insertion of adductor muscle bases ; brachiophores divergent, blade-like 
defining crenulated sockets ; muscle impressions unknown. 

Type material. 

Holotype External and internal moulds of pedicle valve 

(BB 35513a, b) 
Paratypes External and internal moulds of brachial valve 

(BB 35514a, b) 

Incomplete external mould of pedicle valve 

(BB 35515) 

Incomplete internal mould of brachial valve 

(BB 35516) 

Incomplete external and internal moulds of 

pedicle valve (BB 35517a, b) 

Incomplete external and internal moulds of 

brachial valve (BB 35518a, b) 

Incomplete external and internal moulds of 

brachial valve (BB 35519a, b) 

Horizon and localities. Whittery Shales : BB 35516 and BB 35519 from 
exposures about half-way down path going south to valley bottom below Marrington 
Farm (Grid Ref. SO 272967) ; the remainder from lower part of western bank of 
River Camlad 60 yds north-east of Marrington Farm (Grid Ref. SO 272970). 

Discussion. The Reuschella occurring in the Whittery and Hagley Shales are 
related to the early Caradocian R. horderleyensis (s.l.) which is widespread in E. 
Shropshire and Wales (Bancroft 1945:239; Williams 1963:413). However, 
comparisons with R. cf. horderleyensis and R. horderleyensis undulata from the Bala 
area (Williams 1963 : 415-417) show differences exist which merit taxonomic 
recognition. In the outline of the shell and of the ventral muscle scar and its forward 
extension as well as the texture and branching of their radial ornamentation, the 
Shelve specimens are indistinguishable from R. horderleyensis and the Welsh sub- 
species. They do, however, differ from both taxa in being much flatter laterally 
and additionally from R. horderleyensis in the growth of the cardinalia. It seems 



length 


width (mm) 


13-5 


17-0 


13-5 


19-0 


13-5 


- 


13-0 


- 


13-5 


- 


12-5 


- 


4-0 


- 



106 SHELVE DISTRICT 

appropriate, therefore, to recognize the Shelve sample as a new subspecies of R. 
horderleyensis. 

Table 73 

Statistics of length (1) and maximum width (w) of 7 brachial valves of Reuschella horderleyensis 

carinata subsp. no v. 

1 mm (var 1) io-oo (28-5) 

w mm (var w) r3'73 (4 I-I 99) 

r 0-993 

log e l (var log e l) 2-1772 (0-2506) 

logew (var log e w) 2-5206 (0-1976) 

r e 0-995 

a (var a) 0-8879 (0-00155) 

Table 74 

Statistics of length (1) and thickness (th) of 7 brachial valves of Reuschella horderleyensis carinata 

subsp. nov. 

1 mm (var 1) 11-31 (20-841) 
th mm (var th) 1-74 (0-556) 

r o-8n 

a (var a) 0-1633 (0-00182) 

Table 75 

Statistics of length (1) and length of brachiophore bases (lc) in 9 brachial valves of Reuschella 

horderleyensis carinata subsp. nov. 

1 mm (var 1) 12-69 (16-259) 
lc mm (var lc) i-8i (0-196) 

r 0-883 

a (var a) 0-1098 (0-00037) 

Table 76 

Statistics of length (1) and maximum lateral extension (w) of the brachiophore bases of 8 brachial 
valves of Reuschella horderleyensis carinata subsp. nov. 

1 mm (var 1) 1-96 (0-103) 

w mm (var w) 3-99 (0-373) 

r 0-771 

a (var a) 1-9052 (0-24495) 

Family HETERORTHIDAE Schuchert & Cooper 1931 

HETERORTHIS Hall & Clarke 1892 

Heterorthis sp. 

(PI. 17, figs. 10-14) 

Diagnosis. Small, semi-elliptical, piano- to concavo-convex Heterorthis with 
parallel-sided, bilobed ventral muscle scar 42% as long as the pedicle valve and a 
simple, median, plate-like cardinal process. 



ORDOVICIAN BRACHIOPODA 107 

Description. Small, semi-elliptical, piano- to concavo-convex Heterorthis with 
cardinal angles approximating to 90 ° ; pedicle valve about one-tenth as deep as 
long, slightly carinate medially ; brachial valve averaging 80% as long as wide for 
5 valves, plane to gently concave with a slight median sulcus ; ventral and dorsal 
interareas planar, apsacline and anacline respectively, delthyrium open, notothyrium 
covered by chilidium ; radial ornamentation costellate with low rounded ribs 
numbering 5 per mm, 2 mm antero-medially of umbo. 

Teeth small, supported by divergent dental plates ; ventral muscle field parallel- 
sided and elongately bilobed extending forward for an average of 42% the length of 
3 pedicle valves and only slightly narrower than long. 

Cardinal process consisting of a simple median plate with slightly splayed posterior 
surface, embedded in well-developed notothyrial platform, brachiophores plate-like 
acutely divergent, sockets simple without bounding fulcral plates ; adductor scars 
elongately oval impressed on either side of low median ridge, extending anteriorly 
for over half the valve length. 

Figured material. 

length width (mm) 

External and internal moulds of pedicle valve (BB 35423a, b) 4*0 - 

External and internal moulds of brachial valve (BB 35424a, b) 4-5 5-5 

Horizon and locality. Spy Wood Grit exposed 1100 yds NNE of Rorrington 
(Grid Ref. SJ 303015)- 

Discussion. The Heterorthis collected from the Shelve area consist of only 12 
valves from the Spy Wood Grit and one from the Aldress Shales. The largest of 
these specimens was not more than 5 mm long and, although one would expect 
some mature shells even in the small sample available, there is a strong possibility 
that the collections are moulds of young shells. Indeed the undifferentiated cardinal 
process and the relatively modest development of the notothyrial platform suggest 
that this is so ; and until more is known about the age structure of such samples, 
infrageneric recognition is withheld. In that respect the only clue to specific affilia- 
tion is given by the relatively poor development of the lateral diductor lobes in the 
manner of adult H. retrorsistria M'Coy (Williams 1963 : 420) although even that 
character may be merely diagnostic of immaturity. 



TISSINTIA Havlicek 1970 
1970 Tissintia Havlidek : 14. 

When Havlicek proposed Tissintia as a new genus, he rightly emphasized its close 
relationship with Heterorthis. In consideration of the stratigraphic distribution of 
the only species recognized by him as belonging to the genus, he concluded that the 
stock bore many attributes of the prototypic heterorthid (Havlicek 1970 : 9-10). 
The planoconvex to ventribiconvex profile, the bilobed cardinal process and the 
disposition of the brachiophores in the even older T. prototypa confirm his identifica- 
tion of the changes such features underwent to give rise to the characteristics of the 



108 SHELVE DISTRICT 

Caradocian heterorthids. The reflexed costellae along the posterior edges of the 
shells of all heterorthids are especially interesting because well-preserved specimens 
from the Weston Beds show corresponding follicular embayments indenting the 
hinge-lines of both valves. These embayments indicate the presence, at some time 
or another, of backward-projecting setae. Whether functional setae persisted 
much within the cardinal angles is doubtful because the embayments may have 
been incorporated within laterally expanding hinge-lines after they had ceased to 
accommodate setae. 

With regard to the origin of the heterorthids, the existence of so typical a Tissintia 
species as prototypa in the Lower Llanvirn suggests that Paurorthis, the earliest forms 
of which are not much older, may not be ancestral to the Heterorthidae as Havlicek 
(1970 : 16) believed. Indeed as in many other brachiopod groups, proposed lines 
of descent can be no more than tentative guesses until more is known of the 
Tremadoc-Arenig faunas. 



Tissintia prototypa (Williams) emended A. W. 
(PL 17, figs. 15-19 ; PL 18, figs. 1-9, 11) 
1949 Dalmanella prototypa Williams : 168. 

Diagnosis. Subcircular, ventribiconvex Tissintia with fine costellae reflexed 
posteriorly along the hinge-line and numbering 5 per mm, 5 mm antero-medially of 
the dorsal umbo ; brachiophore plates parallel with median ridge with bases extend- 
ing anteriorly for 15% of the length of the brachial valve and 110% of their lateral 
spread ; ventral muscle field elongately bilobed, extending forward for 38% of the 
length of the pedicle valve. 

Description. Subcircular, ventribiconvex to rarely planoconvex Tissintia 
with obtuse cardinal angles ; pedicle valve 21% as deep as long with an evenly 
convex median zone and flattened lateral areas, brachial valve 80% as long as wide 
and about one-tenth as deep as long or rarely planar, shallow median sulcus dying 
out on average 5 mm anterior of dorsal umbo (variance 0-016 for 10 valves), lateral 
areas flattened or slightly concave ; ventral interarea planar, apsacline, longer than 
anacline dorsal interarea, delthyrium and notothyrium open ; radial ornamentation 
fascicostellate commonly 5 per mm, 5 mm antero-medially of the dorsal umbo, 
external branches poorly developed in the first four sectors, costellae reflexed postero- 
laterally so that corresponding follicular embayments indent the internal surfaces of 
the hinge-lines of both valves. 

Teeth small, trigonal, supported by dental plates extending anteriorly for 14% of 
the length of the pedicle valve and diverging for twice their length ; ventral muscle 
field elongately bilobed divided by a pair of fine median ridges extending forward 
from conspicuous pedicle callist ; adductor scars small, elongately oval, submedially 
situated ; diductor scars impressed anteriorly for 38% of the length of the pedicle 
valve but not enclosing adductors, adjustor scars impressed on the median surfaces 
of the dental plates ; mantle canal pattern lemniscate. 



ORDOVICIAN BRACHIOPODA 109 

Cardinal process with bilobed myophore and long shaft on well-developed noto- 
thyrial platform passing anteriorly into median ridge ; brachiophores triangular in 
outline, acutely divergent, brachiophore bases parallel, ankylosed to median ridge 
and extending anteriorly for 15% of the length of the brachial valve, sockets oblique, 
bounded ventrally by brachiophores but not defined laterally by fulcral plates ; 
quadripartite adductor scars 46% as long as the valve ; adjustor scars small oval, 
located on median ridge anterior to notothyrial platform, mantle canal pattern 
lemniscate. 

Figured material. 

Internal mould of brachial valve (BB 35305) 
Internal mould of pedicle valve (BB 35306) 
External mould of brachial valve (BB 35307) 
External mould of brachial valve (BB 35308) 
Internal mould of brachial valve (BB 35309) 
Incomplete internal mould of brachial valve (BB 35310) 
External and internal moulds of pedicle valve (BB 35311a, b) 
External and internal moulds of pedicle valve (BB 35312a, b) 
External and internal moulds of brachial valve (BB 35313a, b) 
External and internal moulds of brachial valve (BB 35314a, b) 
External and internal moulds of brachial valve (BB 35315a, b) 

Horizons and localities. BB 35305-35310 from Lower Llanvirn ashy shales 
exposed in stream 240 yds west of Llwyn Bedw Farm, Llandeilo (Grid Ref. 
SN 653212) ; BB 35311, BB 35314, BB 35315 from Weston Beds exposed 210 yds 
south-east of road crossing Betton Dingle, Lyde (Grid Ref. SJ 317015) ; BB 35313 
from Weston Beds at Cwm Dingle 640 yds south of Little Weston (Grid Ref. 
SO 294978) ; BB 35312 from Weston Beds in road exposure 430 yds south-west of 
Miner's Arms, Priestweston (Grid Ref. SO 291970). 



Tissintia immatura (Williams) emended A. W. 

(PI. 18, figs. 10, 12-15 ; PL 19, figs. 1-5) 

1949 Resserella immatura Williams : 165. 

Diagnosis. Subcircular, planoconvex Tissintia with subcarinate pedicle valve 
31% as deep as long, costellae reflexed posteriorly along the hinge-line and commonly 
numbering 3 per mm 5 mm antero-medially of the dorsal umbo ; brachiophore 
bases normally thickened to extend anteriorly for 17% of the length of the brachial 
valve and 84% of their lateral spread ; ventral muscle field elongately bilobed 
extending forward for 38% of the length of the pedicle valve. 

Description. Subcircular, planoconvex to more rarely ventribiconvex with 
obtuse cardinal angles ; pedicle valve 31% as deep as long with a slightly carinate 



length 


width (mm 


6-0 


7-0 


9-0 


io-o 


7-5 


9-0 


8-o 


9-5 


14-0 


16-0 


14-0 


20-5 


12-5 


15-0 


13-0 


20-0 


13-0 


16-0 


12-0 


14-0 



no SHELVE DISTRICT 

median zone and gently convex lateral areas ; brachial valve 79% as long as wide, 
gently convex or more commonly planar with indistinct sulcus ; ventral interarea 
planar apsacline, longer than anacline dorsal interarea, delthyrium and notothyrium 
open ; radial ornamentation fasciocostellate commonly 3 per mm, 5 mm antero- 
medially of the dorsal umbo, external branches poorly developed in the first three 
sectors, costellae reflexed postero-laterally, follicular embayments rarely indenting 
internal surfaces of hinge-line. 

Teeth small, supported by dental plates extending anteriorly for 17% of the 
length of the pedicle valve and extending laterally for 150% of their length ; ventral 
muscle field elongately oval extending forwards for 38% of the length of the valve 
with splayed diductor scars almost enclosing a pair of small suboval adductor 
impressions separated by a low ridge ; adjustor scar located on dental plates. 

Cardinal process bilobed, differentiated in adult valves into elongate shaft and 
expanded myophore, notothyrial platform well developed passing anteriorly into 
low ridge ; brachiophores triangular in outline with divergent bases extending 
forward for 17% of the length of the valve, becoming massive in adult valves so that 
they are, on average, only 84% as long as their lateral spread ; oblique sockets not 
bounded laterally by fulcral plates ; adductor scars quadripartite extending forward 
for half the length of the valve. 

Figured material. 

External and internal moulds of pedicle valve (BB 35474a, b) 
External and internal moulds of brachial valve (BB 35475a, b) 
External and internal moulds of pedicle valve (BB 35476a, b) 
Incomplete external and internal moulds of brachial valve 
(BB 35477a, b) 

External and internal moulds of brachial valve (BB 35478a, b) 
Internal mould of pedicle valve (BB 35479) 

Horizon and localities. Meadowtown Beds : BB 35474, 35475 from loose 
blocks on rough ground 100 yds south-east of Minicop Farm (Grid Ref. SJ 315018) ; 
BB 35478 from Bed B in measured section in Meadowtown Quarry (Grid Ref. 
SJ 312012) ; BB 35479 from Quinton's Quarry in the field 200 yds north-east of 
Meadowtown Chapel (Grid Ref. SJ 312013). BB 35476, 35477 from exposures in a 
ploughed field along the strike of Betton Beds occurring 180 yds south-west of a well 
in the lane-side from Meadowtown to Castle Ring (Grid Ref. SJ 310009). 

Discussion. Tissintia prototypa was first described from the upper part of the 
Didymograptus bifidus shales of the Llandeilo district (Williams 1953 : 180) and is 
now known from W. Wales and Builth as well as Shropshire, where it occurs abun- 
dantly in the Llanvirn Weston Beds. A good sample from the Weston Beds has been 
statistically compared with a topotypic assemblage. The comparison shows 
(Tables 77-85) that there is no difference between the two samples in parameters of 
the shape of the shell, the disposition and size of internal features, or in the radial 
ornamentation except for a marginally significant delay in the differentiation of the 



mgth 


width (mm) 


8-o 


- 


4-0 


5-o 


9-0 


12-0 


14-0 


- 


10-0 


12-0 


n-5 


- 



ORDOVICIAN BRACHIOPODA in 

3ala costella. This distinction is not considered important enough to merit system- 
atic recognition so that the species may be regarded as a noteworthy homogeneous, 
short-ranging stock especially characteristic of the Lower Llanvirn of Britain. 

The Llandeilo species T. immatura from the Meadowtown Beds is more closely 
related to T. prototypa than was originally thought when the two species were 
erected. Both species are similar in dorsal outline and the relative sizes of internal 
features. They differ however in a number of important characters. In T. immatura 
the ribbing is significantly coarser (p > o-ooi) with a suggestion that externally 
branching ribs are more commonly developed in Sector IV ; the growth in relative 
depth of the pedicle valve is faster (p < o-ooi) ; and the lateral spread of the brachio- 
phore bases is significantly greater (0-05 > p > 0-02). This last difference reflects 
not only the more divergent disposition of the brachiophore bases in T. immatura 
but also their excessive thickening by secondary shell accretion in mature stages of 
growth. 

A small sample of Tissintia from the Betton Beds of Shelve, although strati- 
graphically intermediate between the two species, was identical with T. prototypa 
in the key differences. Rib counts per mm, 5 mm antero-medially of the umbones 
of 2, 3 and 5 brachial valves were 4, 5 and 6 ; the basic statistics for length (1) and 
depth (th) of 11 pedicle valves were: I (var 1) = 7-68 (3-872), th (var th) = 1-53 
(0-27) and r = 0-909 ; and for the length (1) and maximum lateral extension (w) 
of the brachiophore bases in 11 brachial valves they were : I (var 1) = 1-15 (0-137), 
w (var w) = I'll (0-179) an d r = 0-936. Allometry prevailed in both the deepening 
of the pedicle valve and the extension of the brachiophores, but neither growth rates 
nor shape parameters differed significantly from those of T. prototypa. 

The only other species assigned to the genus is T. convergens from the Llandeilo 
of Morocco (Havlicek 1971 : 51), which appears to differ from the British species, 
at least, in being relatively more transverse and in possessing smaller brachiophores 
and longer, more flabellate ventral muscle scars. 



Table 77 

Statistics of length (1) and length of brachiophore bases (lc) in n brachial valves of Tissintia 

prototypa (Williams) from the Weston Beds (A) and the Lower Llanvirn of Llandeilo (B), and of 

T. immatura (Williams) from the Meadowtown Beds (C) 

ABC 

n 21 55 64 
1 mm (var 1) 8-87 (14-326) 8-59 (4-4) 6-03 (14-756) 

rcmm(varlc) 1-4(0-303) 1-17(0-054) 1-05(0-361) 

r 0-941 0-871 0-966 

log e l (var log e l) 2-0986 (0-1674) 2-122 (0-0578) 1-6267 (°'34°4) 

logelc (var logelc) 0-2685(0-1426) 0-1384(0-0387) —0-0903(0-2824) 

r e 0-944 °'875 0-972 

a (var a) 0-9229 (0-00485) 0-8182 (0-00295) 0-9108 (0-00075) 



ii2 SHELVE DISTRICT 

Table 78 

Statistics of length (1) and length of adductor scars from umbo (sc) in n brachial valves of Tissintia 

prototypa (Williams) from the Weston Beds (A) and the Lower Llanvirn of Llandeilo (B), and of 

T. immatura (Williams) from the Meadowtown Beds (C) 

A B C 

n 12 6 34 

1 mm (var 1) 10-43(8-345) 11-18(3-402) 8-08(15-198) 

sc mm (var sc) 4-77 (1-599) 5-13 (0-926) 4-04(2-818) 

r 0-935 0-979 0-984 

a (var a) 0-4377 (0-00242) 0-5217 (0-0028) 0-4306 (0-00018) 

Table 79 

Statistics of length (1) and maximum width (w) of n brachial valves of Tissintia prototypa 

(Williams) from the Weston Beds (A) and the Lower Llanvirn of Llandeilo (B), and of 

T. immatura (Williams) from the Meadowtown Beds (C) 

ABC 

n 27 54 52 

1 mm (var 1) 6-92(18-479) 7-62(9-47) 5-66(11-628) 

w mm (var w) 9-01(29-887) 9-25(10-229) 7-11(15-49) 

r 0-987 0-903 0-99 

logel ( var log e l) 1-771(0-3265) 1-9559(0-1508) 1-5786(0-3097) 

log e w (var log e w) 2-0424(0-3131) 2-1678(0-1129) 1-8275(0-2674) 

r e 0-989 0-91 0-992 

a (var a) 0-9793 (0-00464) 0-8653 (0-00247) 0-9292 (0-00027) 

Table 80 

Statistics of length (1) and thickness (th) of n pedicle valves of Tissintia prototypa (Williams) from 
the Weston Beds (A) and the Lower Llanvirn of Llandeilo (B), and of T. immatura (Williams) 

from the Meadowtown Beds (C) 

ABC 
n 14 31 49 

1 mm (var 1) 10-81(18-219) 7 - 88 (3-719) 7-18(12-06) 

th mm (var th) 2-59 (1-299) I- 45 (° -I 7) 2-19 (0-925) 

r 0-895 0-800 0-932 

a (var a) 0-267 (0-00118) 0-2138 (0-00057) 0-277 (0-00022) 

Table 81 

Statistics of length (1) and length of dental plates (dl) in n pedicle valves of Tissintia prototypa 

(Williams) from the Weston Beds (A) and the Lower Llanvirn of Llandeilo (B) , and of T. immatura 

(Williams) from the Meadowtown Beds (C) 

ABC 

n 21 27 53 

1 mm (var 1) 11-45(9-102) 8-36(3-282) 6-92(14-129) 

dl mm (var dl) 1-68 (0-328) 1-06 (0-055) i'i8 (0-341) 

r 0-909 0-624 0946 

logel (var log e l) 2-4044(0-0671) 2-0999(0-0461) 1-8049(0-2586) 

logedl (var logedl) 0-4614(0-1103) 0-0297(0-0484) 0-0522(0-2205) 

r e 0-914 0-627 0-952 

a (var a) 1-2822 (0-0143) 1-0246 (0-02548) 0-9234 (0-00157) 



ordovician brachiopoda 113 

Table 82 

Statistics of length (1) and length of ventral muscle scar (sc) in n pedicle valves of Tissintia 

prototypa (Williams) from the Weston Beds (A) and the Lower Llanvirn of Llandeilo (B), and of 

T. immatura (Williams) from the Meadowtown Beds (C) 

ABC 

n 20 17 43 

lmm(varl) 11-45 (9-581) 9-51(2-327) 7-5(13-902) 

sc mm (var sc) 4-11 (i-88i) 3-31 (0-432) 2-84 (2-991) 

r 0-934 0-729 0-96 

log e l (var log e l) 2-4023(0-0706) 2-2392(0-0255) 1-9048(0-2208) 

logesc (var log e sc) 1-3606(0-1056) 1-1762(0-0389) 0-8848(0-316) 

r e 0-937 0-73 0-967 

a (var a) 1-2231(0-01008) 1-2351(0-04747) 1-1963(0-00229) 



Table 83 

Statistics of length (1) and maximum lateral extension (w) of the brachiophore bases of n brachial 
valves of Tissintia prototypa (Williams) from the Lower Llanvirn of Llandeilo (A) and T. imma- 
tura (Williams) from the Meadowtown Beds (B) 

A B 

n 61 67 

1 mm (var 1) 1-22 (0-069) I-0 9 (0-382) 

w mm (var w) i-i (0-065) I-2 9 (0-44 1 ) 

r 0-895 0-927 

logel ( var logel) 0-1729(0-0457) -0-0527(0-2783) 

logew (var logew) 0-0707 (0-052) 0-1372 (0-235) 

r e 0-9 0-935 

a (var a) 1-0674 (0-00367) 0-919 (0-00163) 



Table 84 

Proportions of the earlier insertion of 10 costellae relative to 10 others (with the number of bra- 
chial valves in which both costellae originated at the same growth stage in brackets) for Tissintia 
prototypa (Williams) from the Weston Beds (A) and the lower Llanvirn of Llandeilo (B) and for 
Tissintia immatura (Williams) from the Meadowtown Beds (C) 





A 


B 


C 


lal ) IB 


i/7 (1) 


0/6 (2) 


i/4 (2) 


IB ) la° 


8/8 


10/13 (1) 


5/9 (1) 


2B ) 2a 


9/9 


15/17 (1) 


11/12 (2) 


2C ) 2a° 


i/3 


2/" (3) 


0/4 (1) 


2a.l ) 2B 


i/3 (5) 


o/5 (7) 


2/8 (3) 


3ala ) 3a 


i/4 (4) 


0/14 


3/14 (1) 


3£ ) 3a 


2/7 (2) 


0/15 


1/12 (1) 


3ala ) 2a 


8/8 


3/10 (2) 


9/12 


46 ) 4 b° 


8/8 


3/3 (2) 


i/4 (3) 



4&P ) 4BI 0/2 1/3 2/3 



ii 4 SHELVE DISTRICT 

Table 85 

The distribution of brachial valves of Tissintia prototypa (Williams) from the Weston Beds (A) 
and the Lower Llanvirn of Llandeilo (B), and of Tissintia immatura (Williams) from the Meadow- 
town Beds (C), with 3-6 costellae per mm, 5mm antero-medially of the dorsal umbones 

costellae per mm 
3 4 5 6 

A 1 4 11 o 

B 2 18 27 4 

C 12 6 1 o 

Family LINOPORELLIDAE Schuchert & Cooper 1931 

SALOPIA Williams 1955 

Salopia cf. salteri (Davidson) 

(PI. 19, figs. 10, 11) 

An incomplete mould of a brachial valve (BB 35426a, b) is the only record of 
Salopia in Spy Wood Grit. The specimen, which was recovered from the outcrops 
1 100 yds NNE of Rorrington (Grid Ref. SJ 303015), indicates that the convex and 
gently sulcate valve was about 5 mm long and was ornamented by fine multicostellae 
with a density of 6 per mm at the antero-medial margin. Although the mould is 
broken posteriorly it is possible to determine that the cardinalia included a pair of 
widely placed brachiophore bases bounding a low notothyrial platform supporting 
the linear shaft of a cardinal process. The notothyrial platform passes anteriorly 
into a broad median ridge with adductor impressions on either side. 

In respect of its preserved features, the valve is indistinguishable from that of 
Salopia salteri (Davidson) as illustrated in Whittington & Williams (1955, pi. 38, 
figs. 45, 46) and provisionally, at least, may be assigned to that species. 

Salopia sp. 
(PI. 19, figs. 6-9) 

In contrast to the Spy Wood representative of Salopia, the three congeneric moulds 
from the Whittery Shales, exposed in Whittery Quarry (BB 35428a, b, BB 35429a, b) 
at the south end of Whittery Wood near Chirbury (Grid Ref. SO 275981) and in the 
stream (BB 35430a, b) at the north end of Spring Coppice 865 yds south-east of 
Hockleton Bridge (Grid Ref. SO 279997), are less decisively related to 5. salteri. The 
two pedicle valves represented by the moulds BB 35429 and 35430 were strongly 
convex, about one-quarter as deep as long, rectimarginate, with a high apsacline 
interarea. The widely placed dental plates were more nearly parallel in the smaller 
valve than the larger although the muscle fields of both, consisting of a pair of lobate 
diductor scars flanking a narrow median adductor track, were impressed beyond the 
limits of the dental plates and extended forwards for up to one-third the length of 
the valve. The ribbing was too poorly preserved on external moulds to indicate 
more than a finely costellate condition. However, costellae ornamenting the brachial 
valve (BB 35428), although occurring with a density comparable with that of 



ORDOVICIAN BRACHIOPODA 115 

5. salteri of 5 per mm, 5 mm antero-medially of the umbo, are sharper and less uni- 
formly multicostellate. This difference is the only important one observed because 
the apparent greater convexity and sulcation are attributable to post-diagenetic 
deformation of the specimen. When more is known about ribbing variation in 
Salopia the difference may prove to have no systematic status ; meanwhile it seems 
advisable not to allocate the specimens to any described species. 

Suborder CLITAMBONITIDINA Opik 1934 

Superfamily GONAMBONITACEA Schuchert & Cooper 1931 

Family GONAMBONITIDAE Schuchert & Cooper 1931 

Subfamily ANOMALORTHINAE Ulrich & Cooper 1936 

OSLOGONITES Opik 1939 

Oslogonites ? sp. 

(PI. 19, figs. 12, 13) 

An external and incomplete internal mould of a brachial valve (BB 35326a, b) 
about 2 mm long and 3-5 mm wide, from the Mytton Flags exposed along the side 
of the footpath at road level, 1100 yds NNW of Wood House (Grid Ref. SJ 338003), 
is believed to represent the remains of an immature specimen of Oslogonites Opik. 
The valve was semicircular in outline and evenly concave in profile with acute 
cardinal angles. It was ornamented by intercalated costellae, numbering 6 per mm 
anteriorly, with every fourth or fifth thickened to delineate a series of sectors in an 
unequally parvicostellate fashion. Enough of the internal mould is preserved to 
show that the socket ridges were acutely disposed to the hinge-line to define narrow 
sockets and joined medially in a small ridge which was not differentiated into a 
cardinal process. There was no notothyrial platform nor median ridge, but a pair 
of curved septa occurred in the area normally occupied by adductor scars. 

The concavity of the brachial valve and the unequally parvicostellate nature of its 
radial ornamentation suggest either a strophomenidine or a clitambonitidine affinity. 
However, the disposition of the socket ridges favour its identification as a clitam- 
bonitidine ; and since the anomalous lack of a notothyrial platform may be attribut- 
able to the immaturity of the valve, the specimen is best assigned to Oslogonites. In 
fact the Shelve brachial valve probably resembled that of Oslogonites costellatus 
(Opik 1939 : 134) from the Arenigian expansus Shale of Norway except that it was 
more concave and its parvicostellate ornamentation more strongly differentiated 
into sectors. 

Family KULLERVOIDAE Opik 1934 

KULLERVO Opik 1932 

Kullervo sp. 

(PI. 19, figs. 14, 15) 

The only kullervoidid represented in the Whittard collection consists of a 
fractured internal mould and an incomplete complementary external mould of a 



n6 SHELVE DISTRICT 

pedicle valve (BB 35340a, b) from the Meadowtown Beds exposed in the lane to 
Waitchley 140 yds north of Meadowtown Chapel (Grid Ref. SJ 311014). These 
meagre fragments indicate that the valve was about 4-0 mm long, 5-5 mm wide and 
1 -5 mm deep with a high apsacline interarea, slightly acute cardinal angles and a 
faint antero-medial sulcus. Radial ornamentation consisted of about 20 rounded 
costae and a few costellae arising by branching with 3 ribs per mm at the antero- 
lateral margin. The ribs and interspaces were crossed by strong, outwardly de- 
flected lamellae. This concentric ornamentation is regularly arranged with a 
density of 4 lamellae per mm at the valve margin and breaks the continuity of the 
ribs. The valve interior was dominated by a spondylium supported by a median 
septum extending forward for 1-5 mm. 

Consideration of the valve profile and outline, ornamentation, and the nature of the 
spondylial support, suggests that the specimen is best assigned to Kullervo. Indeed, 
sufficient radial ornamentation is preserved to prompt comparison with one of the 
earliest recorded species, K. panderi (Opik 1934 : 164) from the C 2 horizon of 
Estonia, because there is no strong differentiation between the postero-lateral and 
medial sectors of radial ornamentation as in later species. However, until more 
material is obtained to assess the validity of this comparison, only a generic identi- 
fication is appropriate. 



Suborder TRIPLESIIDINA Moore 1952 

Superfamily TRIPLESIACEA Schuchert 1913 

Family TRIPLESIIDAE Schuchert 1913 

TRIPLESIA Hall 1859 

Triplesia sp. 

(PI. 19, fig. 16) 

An incomplete internal mould (BB 35408), with fragments of the shell adhering, 
of the genus Triplesia have been recovered from the Whittery Shales exposed in the 
stream at the north end of Spring Coppice, 865 yds south-east of Hockleton Bridge 
(Grid Ref. SO 279997). The mould is that of a pedicle valve which was 12-5 mm 
long, 10 mm wide and 2-5 mm deep laterally. Its outline was subquadrate with 
rounded obtuse cardinal angles and it was indented medially by a strong flat- 
bottomed sulcus, about half as wide as the maximum width of the valve, which 
was sharply bent into a right angle to project dorsally as a tongue about one- 
third as long as the ventral length of the valve. The apsacline interarea, which 
was curved, bore a medially folded pseudodeltidium, and the external surface bore 
only fine overlapping lamellae numbering about 14 per mm anteriorly. Internally 
the dental plates were acutely divergent, extending forward for about one-fifth the 
length of the valve, while the diductor scars were lightly impressed for over half the 
length of the valve on either side of a low median ridge. 

The Whittery Triplesia with its flat-bottomed sulcus and sharply bent, dorsally 
projecting tongue is immediately distinguishable from penecontemporaneous 



ORDOVICIAN BRACHIOPODA 117 

European and African species, such as T. simplex from the Ktaoua Formation of 
Morocco (Havlicek 1971 : 66) and T. deformata (Barrande 1879) from the Zahofany 
Formation of Czechoslovakia (Havlicek 1950 : 89), with their gently rounded non- 
geniculate sulci. It is, however, similar to T. subcarinata (Cooper 1956 : 538) from 
the Lebanon Formation of Tennessee ; but systematic consideration of this apparent 
relationship must await the information provided by more material. 

BICUSPINA Havlicek 1950 

Bicuspina subquadrata sp. nov. 

(PI. 19, figs. 17-19 ; PL 20, figs. 1, 2) 

Diagnosis. Subquadrate Bicuspina, with a dorsal fold about 44% as wide as the 
brachial valve length, usually bearing a median costa which rarely branches within 
5 mm of the umbo, finely lamellose with 7 lamellae developed between 5 and 6 mm 
antero-medially of the umbo. 

Description. Medium-sized, dorsibiconvex, subquadrate Bicuspina with 
rounded, obtuse cardinal angles ; pedicle valve 25% as deep as long with flat- 
bottomed median sulcus bounded by parallel sides and evenly convex flanks ; brachial 
valve 65% as long as wide and 37% as deep as long with a strong median fold 44% 
as wide as the valve is long ; ventral interarea curved apsacline, pseudodeltidium 
with narrow median fold, dorsal interarea vestigial ; radial ornamentation of 
rounded ribs with a mean wavelength of 0-44 mm (variance 0-02), 5 mm anterior of 
the umbones of 15 brachial valves — a median costa (flanked by 2 laterals) occurs on 
folds of 11 out of 16 brachial valves and bifurcates within 5 mm of the umbo in 
only 1 valve — 9 to 15 ribs occur on the flanks of brachial valves between 5 and 
7-5 mm long ; concentric ornamentation of delicate overlapping lamellae with 6 or 
7 occurring between 5 and 6 mm anteriorly of the umbones of 2 and 7 brachial valves 
respectively. 

Pedicle tube relatively short, averaging 12% the length of the pedicle valve and 
with slightly curved pedicle passage ; rounded teeth supported by acutely divergent 
dental plates 60% as long as their anterior separation and extending forward for 
17% of the length of the valve ; suboval muscle field well impressed in late adult 
stages of growth only, extending forward for 48% of the length of the pedicle valve 
and consisting of a short median adductor and a pair of ventral adjustors situated 
on either side of the internal opening of the pedicle tube, which is flanked by long 
diductor scars. 

Cardinalia consisting of forked, posteriorly curving cardinal process, and short, 
acutely divergent, pointed brachiophores ; sockets disposed parallel to the hinge-line, 
extending laterally for 33% of the length of the valve, not restricted distally by 
ridges of secondary shell. 

Type material. 

length width (mm) 
Holotype Internal and incomplete external moulds of 

pedicle valve (BB 35480a, b) 18-0 20-0 



u8 SHELVE DISTRICT 

length width (mm) 
Paratypes External and internal moulds of brachial valve 

(BB 35481a, b) n-5 

External and incomplete internal moulds of 
pedicle valve (BB 35482a, b) 12-0 16-5 

Internal mould of brachial valve (BB 35483) 14-5 22-5 

Horizon and localities. Whittery Shales : BB 35480 from Tuffaceous Shale 
exposed in the lower part of the western bank of the River Camlad 60 yds north-east 
of Marrington Farm (Grid Ref. SO 272970) ; other specimens from Whittery 
Quarry at the south end of Whittery Wood near Chirbury (Grid Ref. SO 275981). 

Bicuspina modesta sp. nov. 

(PI. 20, figs. 3-8) 

Diagnosis. Subquadrate Bicuspina with a dorsal fold about half as wide as the 
valve length, commonly bearing a median costa which usually branches within 5 mm 
of umbo, lamellose with 5 lamellae developed between 5 and 6 mm antero-medially. 

Description. Small, biconvex, subquadrate Bicuspina with obtuse or, more 
rarely, orthogonal cardinal angles ; pedicle valve 24% as deep as long with a flat- 
bottomed median sulcus bounded by parallel sides and evenly convex flanks, brachial 
valve 75% as l° n g as wide and 35% as deep as long with a splayed median fold 50% 
as wide as the valve is long ; ventral interarea curved apsaline, pseudodeltidium 
with narrow median fold, dorsal interarea vestigial ; radial ornamentation of 
rounded ribs with a mean wavelength of 0-5 mm (variance 0-08), 5 mm anterior of 
the umbones of 27 brachial valves, median costa (flanked by 2 laterals) occurs on 
folds of 34 out of 37 brachial valves and bifurcates within 5 mm of the umbones of 
19 valves ; 11 to 17 ribs occur on the flanks of brachial valves between 5 and 7-5 mm 
long ; concentric ornamentation of overlapping lamellae with 4 or 5 occurring 
between 5 and 6 mm anteriorly of the umbones of 2 brachial valves in each case. 

Pedicle tube and passage averaging 11% of the length of the pedicle valve ; small 
teeth supported by acutely divergent dental plates extending forward for 68% of 
their anterior separation and 15% as long as the valve ; muscle field poorly impressed 
extending forward for 55% of the length of the pedicle valve. 

Cardinalia consisting of forked posteriorly curving cardinal process and short, 
acutely divergent, pointed brachiophores ; sockets disposed parallel to the hinge- 
line extending laterally for 26% of the length of the valve, usually defined distally 
by ridges of secondary shell. 

Type material. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 35510a, b) io-o 12-0 

Paratypes External and internal moulds of brachial valve 

(BB 35511a, b) 7-0 8-5 

External and internal moulds of brachial valve 

(BB 35512a, b) 9-5 n-5 



ORDOVICIAN BRACHIOPODA 119 

Horizon and locality. Spy Wood Grit exposed 1100 yds north-east of 
Rorrington (Grid Ref. SJ 303015). 

Discussion. Bicuspina is well represented in the Spy Wood Grit and the 
Whittery Shales by two closely related forms which are, nonetheless, distinguishable 
in the relative size of the median fold and sulcus and in ornamentation. As com- 
parison of growth axes demonstrated (Tables 86-93), the two species are alike in the 
outline and depth of the valves as well as in the relative growth of dental plates and 
cardinalia. The dorsal fold of B. subquadrata, however, is relatively narrower than 
that of B. modesta, and a /-test indicates that the difference was maintained during 
shell growth at a significant level (o-oi > p > o-ooi). This difference may have 
influenced the incidence of ribs on the folds because a median costa is more fre- 
quently absent or, if present, more likely to have bifurcated in later stages of shell 
growth in B. subquadrata than in B. modesta. Concentric lamellae of the former 
species are also more delicate and more closely spaced than those of B. modesta. 

The Shelve species may be ancestral to B. spiriferoides (M'Coy) which is widely 
distributed in late Soudleyan and Longvillian rocks of Shropshire and N. Wales. 
Lack of suitable material has precluded study of the variability of the radial ornament 
and especially the development of a dorsal median costa (Williams 1963 : 426). B. 
spiriferoides, however, is distinct in its relatively greater width, its rectangular to 
acute cardinal angles, and its relatively long dental plates and ventral muscle field. 

The Bicuspina described by Havlicek (1950 : 88-90) from the late Llandeilian- 
early Caradocian (Chrustenice and Lodenice substages) rocks of Czechoslovakia also 
differ in a number of respects but especially in the relative coarseness of ribbing in 
B. cava (Barrande) and the lack of fold and sulcus in B. multicostellata Havlicek. 



Table 86 

Statistics of length (1) and maximum width (w) of n brachial valves of Bicuspina subquadrata 
sp. nov. (A) and Bicuspina modesta sp. nov. (B) 

A B 

n 7 34 

1 mm (var 1) 9-10 (10-037) °"55 (8-024) 

w mm (var w) 13-94 (19-843) 8-70 (13-113) 

r 0-986 0-966 

a (vara) 1-4061 (0-01130) 1-2783 (0-03384) 

Table 87 

Statistics of length (1) and maximum depth (th) of n brachial valves of Bicuspina subquadrata 
sp. nov. (A) and Bicuspina modesta sp. nov. (B) 





A 


B 


n 

1 mm (var 1) 

th mm (var th) 

r 
a (var a) 


10 

8-13 (4*447) 

3-00 (1-315) 

0-916 

o-5439 (0-00595) 


33 

6-70 (7-165) 
2-36 (1-033) 
0-869 
o-3797 (0-00114) 



120 shelve district 

Table 88 

Statistics of length (1) and maximum width of folds (f) in n brachial valves of Bicuspina sub- 
quadrata sp. nov. (A) and Bicuspina modesta sp. nov. (B) 

A B 

n io 36 

1 mm (var 1) 8-55 (5-660) 6-41 (3757) 

f mm (var f) 379 (2-465) 3-18 (0-747) 

r 0-9537 0-794 

log e l (var log e l) 2-1086 (0-0746) 1-8143 (0-0875) 

logef (var logef) 1-2533 (0-1583) 1-1205 (0-0714) 

r e 0-958 o-8ooo 

a (var a) 1-4567 (0-02400) 0-9033 (0-00886) 

Table 89 

Statistics of length (1) and maximum depth (th) of n pedicle valves of Bicuspina subquadrata 
sp. nov. (A) and Bicuspina modesta sp. nov. (B) 

A B 

n 13 19 

1 mm (var 1) 9-82 (14-945) 7" 12 (2-352) 

tn mm (var th) 2-46 (1-382) 1-68 (0-335) 

r 0-949 0-807 

a (var a) 0-3042 (0-00083) °"3774 (0-00292) 

Table 90 

Statistics of length (1) and length of dental plates (dl) in n pedicle valves of Bicuspina subquadrata 
sp. nov. (A) and Bicuspina modesta sp. nov. (B) 

A B 

n 12 32 

1 mm (var 1) 10-47 (i37 I 3) 7-33 (3-207) 

dl mm (var dl) 1-77 (0-858) 1-13 (0-092) 

_£_ 0-925 0-524 

logel (var logel) 2-2903(0-1178) 1-9637(0-0578) 

logedl (var log e dl) 0-4476 (0-2430) 0-0916 (0-0690) 

r e 0-936 0-533 

a (var a) I- 4363 (0-02974) 1-0927 (0-02640) 

Table 91 

Statistics of length (dl) and anterior separation (dw) of dental plates in n pedicle valves of 
Bicuspina subquadrata sp. nov. (A) and Bicuspina modesta sp. nov. (B) 

A B 

n 12 33 

dl mm (var dl) 1-59 (0-444) I-I 5 (0-076) 

dw mm (var dw) 2-66 (1-284) I- 68 (0-122) 

r 0-905 0-647 

a (var a) 1-7000 (0-05245) 1-2679 (0-03014) 



ordovician brachiopoda 121 

Table 92 

Statistics of length (1) and length (tl) of pedicle tubes in n pedicle valves of Bicuspina subquadrata 
sp. nov (A) and Bicuspina modesta sp. nov. (B) 





A 


B 


n 

1 mm (var 1) 

tl mm (var tl) 

r 


IO 

u-39 (11-677) 
1-41 (o-668) 
0-865 


24 
7-54 (3-423) 
0-85 (0-215) 

o-434 


logel (var logel) 


2-3897 (0-0862) 


1-9907 (0-0584) 


logetl (var logetl) 
a (var a) 


0-1988 (0-2896) 

0-887 

1-8330 (0-10570) 


0-2927 (0-2603) 
o-457 

2-III2 (0-16440) 



Table 93 

Statistics of length (1) and lateral extension of sockets (br) in n brachial valves of Bicuspina 
subquadrata sp. nov. (A) and Bicuspina modesta sp. nov. (B) 

A B 

n 7 17 

1 mm (var 1) 10-21(10-858) 7-46(8-487) 

br mm (var br) 3-37 (0-930) 1-95 (0-596) 

r 0-887 0-905 

a (var a) 0-2927 (0-00365) 0-2650 (0-00084) 



CAEROPLECIA gen. nov. 

Name. A triplesiid with a concentric ornament like the end threads (Gk. KaZpos) 
in a loom. 

Diagnosis. Subcircular, dorsibiconvex, uniplicate triplesiids with a strongly 
convex brachial valve and a gently convex to flat pedicle valve, bearing respectively 
a flat-topped fold and complementary sulcus arising anterior of the umbo ; ventral 
interarea short, apsacline, pseudodeltidium narrowly folded medially, foramen 
supra-apical, dorsal interarea linear ; radial ornamentation of coarse costae and 
branching costellae originating during later stages of shell growth, concentric 
ornamentation of delicate rounded ridges (fila). 

Ventral interior with variably developed pedicle tube, small teeth supported by 
short acutely divergent dental plates, muscle scars broadly bilobed with low median 
ridge dividing narrow adductor band ; mantle canal pattern unknown. 

Dorsal interior with short forked cardinal process, divergent rod-like brachio- 
phores, and small oblique sockets ; fine median ridge dividing quadripartite adductor 
scars with anterior pair larger than posterior ; mantle canal pattern unknown. 

Type species. Caeroplecia plicata sp. nov. from the Whittery Shales, Shropshire. 

Discussion. The new genus is unique among triplesiaceans in being concentri- 
cally ornamented by fine, rounded ridges instead of overlapping lamellae. The 
delayed development of radial ornamentation, the subcircular outline and the flatfish 
pedicle valve together also distinguish Caeroplecia from all other triplesiid genera. 



122 SHELVE DISTRICT 

Caeroplecia plicata sp. nov. 
(PI. 20, figs. 14-16 ; PI. 21, figs. i-6, 8 ; Text-fig. 9) 

Diagnosis. Subcircular Caeroplecia with a brachial valve about 80% as long as 
wide, and low rounded costellae with a wavelength of 0-6 mm, 5 mm antero-medially 
of the dorsal umbo, first appearing on shells more than 3 mm long and commonly 
numbering 6 on the fold and 5 to 9 on the lateral slopes. 

Description. Dorsibiconvex, subcircular Caeroplecia with obtusely rounded 
cardinal angles ; pedicle valve 18% as deep as long with flat-bottomed shallow median 
sulcus and flattened lateral slopes ; brachial valve 80% as long as wide and 30% as 
deep as long with flat median fold, 43% as wide as the length of the valve, originating 
2-5 mm anterior of umbo (average for 4 brachial valves) ; ventral interarea short 
apsacline, pseudodeltidium narrowly folded medially, dorsal interarea linear ; 
concentric ornamentation of 8 to 10 fila per mm between 5 and 6 mm anterior of 
dorsal umbo, radial ornamentation not developed on immature shells up to an average 
of 3-2 mm long (variance 2-233 f° r I0 brachial valves), thereafter rounded costae and 
implanted costellae occur, having a mean wavelength of 0-63 mm, 5 mm anterior 
of the umbones of 4 brachial valves with 5, 6, 7 and 8 ribs on the folds of 2, 3, 1 and 
2 brachial valves between 9 and 12 mm long and 5 to 9 ribs on the lateral slopes. 

Pedicle tube weakly developed, less than 7% as long as the valve ; teeth supported 
by acutely divergent dental plates extending anteriorly for a mean of 11 -8% of the 
length of 6 pedicle valves (variance 5-86) ; muscle field faintly impressed but broadly 
bilobed, extending anteriorly for about one-third the length of the valve, with splayed 
diductor scars not enclosing a narrow parallel-sided adductor track divided by a fine 
median ridge. 

Cardinalia consisting of slightly recurved forked cardinal process, short brachio- 
phores and obliquely placed sockets arranged along the hinge-line for an average of 
24% of the length of 5 brachial valves ; adductor scars quadripartite, lightly im- 
pressed about a fine median ridge. 



- foramen 




cardinal process 



pedicle tube 

socket ridge 
dental plate 



\ run /, 

& mi 



l 1 \ ' 



gonadal sac 
- muscle scar 
-mantle canal 




A B 

Fig. 9. Diagrammatic views of (A) the ventral and (B) the dorsal interiors of 

Caeroplecia. 



ordovician brachiopoda 123 

Type material. 

length width (mm) 

Holotype External and internal moulds of pedicle valve 

(BB 35342a, b) 9-3 n-2 

Paratypes Partly exfoliated pedicle valve with internal 

mould (BB 35343a, b) 12-0 

Internal mould of pedicle valve (BB 35344) n*5 14-5 

Incomplete external and internal moulds of 

brachial valve (BB 35345a, b) - - 

External and internal moulds of pedicle valve 

(BB 35346a, b) io-o 

Incomplete external and internal moulds of 

pedicle valve (BB 35347a, b) - - 

External and internal moulds of brachial valve 

(BB 35348a, b, c) 12-5 13-5 

External and internal moulds of brachial valve 

(BB 35349a, b) 12-2 15-5 

External and internal moulds of brachial valve 

(BB 35350a, b) 6-5 7-5 

Incomplete external and internal moulds of 

brachial valve (BB 35351a, b) - - 

Type horizon and localities. Whittery Shales : BB 35343 and 35347 from 
exposures in the stream at the north end of Spring Coppice, 865 yds south-east of 
Hockleton Bridge near Chirbury (Grid Ref. SO 279997) '> BB 35350 from exposures 
in the lower part of the west bank of the River Camlad 60 yds north-east of Marring- 
ton Farm (Grid Ref. SO 272970) ; all other specimens from Whittery Quarry at 
the south end of Whittery Wood near Chirbury (Grid Ref. SO 275981). 

Discussion. One Welsh species, Oxoplecia mutabilis Williams (in Whittington & 
Williams 1955 : 411), and possibly two American forms, the Porterfield 0. gibbosa 
and the Trenton 0. pennsylvanica (Cooper 1956 : 543, 553) can be assigned to 
Caeroplecia, although the presence in the two latter species of the ridge-like concentric 
ornamentation so characteristic of the genus has still to be demonstrated. However, 
all three differ in the number of ribs on the fold, C. mutabilis having a greater number 
(6 to 12, usually 9) and the American species significantly fewer (3 or 4) . C. mutabilis, 
which is probably the most closely related, also differs in being relatively wider and 
in having grown to lengths of 7 or 8 mm before comparatively fine costae and 
costellae, with a wavelength of 0-3 to 0-5 mm, appear. 

Table 94 

Statistics of length (1) and maximum width (w) of 7 brachial valves of Caeroplecia plicata gen. et 

sp. nov. 

1 mm (var 1) 9-81 (10-558) 

w mm (var w) 12-31 (17-092) 

r 0-995 

a (var a) 1-2724 (0-00311) 



i2 4 shelve district 

Table 95 

Statistics of length (1) and maximum depth (th) of 7 brachial valves of Caeroplecia plicata gen. et 

sp. nov. 

] mm (var 1) 9-81 (10-558) 

th mm (var th) 2-89 (2-381) 

r 0-865 

a (var a) °"474° (0-01136) 



Table 96 

Statistics of length (1) and maximum width of the folds (f) in 6 brachial valves of Caeroplecia 

plicata gen. et sp. nov. 

1 mm (var 1) 9-48 (io-686) 

I mm (var f) 4-12 (2-322) 

r 0-952 

a (var a) 0-4661 (0-00509) 



Table 97 

Statistics of length (1) and maximum depth (th) of 10 pedicle valves of Caeroplecia plicata gen. et 

sp. nov. 

1 mm (var 1) 11-45 (5-679) 

tB mm (var th) 2-02 (0-409) 

r 0-636 

a (var a) 0-2683 (0-00535) 



OXOPLECIA Wilson 1913 
191 3 Oxoplecia Wilson : 81. 

The distinction between Oxoplecia and Bicuspina is finely drawn. Ignoring the 
variability normally inherent to the development of a fold and sulcus, and radial 
and concentric ornamentation, the difference ultimately rests on the spiriferoid 
outline and pedicle tube of Bicuspina (Wright in Williams et al. 1965 : H358). 
These features are likely to have arisen by a lateral acceleration in the growth of the 
cardinal margins and secondary shell accretion around the internal opening of the 
pedicle foramen respectively, and were not always equally developed in stratigraphi- 
cally earlier stocks. Thus Llandeilian Bicuspina of Czechoslovakia (Havlicek 1950 : 
18-19) are strongly orthoid in outline, and contemporaraneous Oxoplecia from S.W. 
Wales have incipient pedicle tubes (R. Addison, pers. comm.). In effect it is likely 
that members of the earlier-occurring and longer-ranging cosmopolitan Oxoplecia 
were ancestral to Bicuspina, which appears to be restricted to the Llandeilian and 
Caradocian rocks of Europe. It is even possible that the species described below is 
more closely related to Bicuspina than contemporaneous American Oxoplecia which 
are typically elliptical in outline, strongly uniplicate, lamellose and coarsely costellate. 



ORDOVICIAN BRACHIOPODA 125 

Oxoplecia cf. nantensis MacGregor 
(PI. 20, figs. 9-13) 
1961 Oxoplecia nantensis MacGregor : 196. 

Diagnosis. Subcircular, subequally biconvex Oxoplecia with a variably 
developed dorsal fold over half as wide as the brachial valve length, and fine radial 
ornament consisting of ribs averaging 0-3 mm in wavelength, 5 mm antero-medially 
of dorsal umbo. 

Description. Small, subequally biconvex Oxoplecia with obtuse cardinal angles 
and a hinge-line about three-quarters the maximum width of the shell ; pedicle 
valve 23% as deep as long with rounded variably defined sulcus, brachial valve 84% 
as long as wide and 29% as deep as long with a broadly rounded fold 57% as wide 
as the brachial valve is long ; ventral interarea curved apsacline, pseudodeltidium 
with narrow median fold, dorsal interarea linear ; up to 10 rounded branching ribs, 
usually not including a median costa and with a mean wavelength of 0-31 mm, 
5 mm anterior of the umbones, occur on the folds and 10 to 12 ribs on the flanks of 
5 brachial valves less than 10 mm long ; concentric ornamentation of overlapping 
lamellae numbering about 7 between 5 and 6 mm anterior of the dorsal umbo. 

Pedicle passage 5% as long as the pedicle valve, rarely with incipient traces of 
pedicle tube, small teeth supported by acutely divergent dental plates extending 
forward for 11% the length of the valve ; muscle field unknown. 

Cardinalia consisting of deeply bilobed cardinal process and short brachiophores, 
simple sockets arranged parallel to the hinge-line and extending laterally for almost 
one-quarter the length of the valve. 

Figured material. 

length width (mm) 
External and internal moulds of pedicle valve (BB 35320a, b) 8-0 8-5 

Incomplete internal mould of brachial valve (BB 35321) 6-0 8-0 

External and internal moulds of brachial valve (BB 35322a, b) 8-5 9-5 

Incomplete internal mould of brachial valve (BB 35323) - 

Horizon and localities. Meadowtown Beds : BB 35322 from exposures along 
the lane to Lower Ridge from bench mark 754, Little Weston (Grid Ref. SO 293984) ; 
the remaining specimens from temporary exposures at the side of a cart-track near 
Waitchley (Grid Ref. SJ 311018). 

Discussion. Oxoplecia is fairly widely distributed throughout the Lower 
Llandeilo rocks of Wales and Shropshire but is never common in any one locality so 
that data about its morphological variability are not abundant. It is, however, 
immediately distinguishable from penecontemporaneous Oxoplecia from other 
regions in the variable development of the fold and sulcus and especially in the 
fineness of the radial ornamentation. This is true even of the most closely com- 
parable American species, the Chazyan Oxoplecia costellata (Cooper 1956 : 540), 
which further differs in the late development of fold and sulcus. The few Oxoplecia 
species reported from Ordovician successions of Eurasia also differ in other attri- 
butes. Oxoplecia sibirica, from the Mangaseya Stage of the Siberian platform 



126 SHELVE DISTRICT 

(Nikiforova & Andreeva 1961 : 200), is additionally Onychoplecia-like in outline 
and even more finely ribbed ; while the species described by Opik (1930 : 200) as 
'Cliftonia' dorsata Hisinger may prove to be a Bicuspina (Roomusoks 1970 : 121) 
rather than an Oxoplecia. 

The Shelve specimens have been assigned to Oxoplecia nantensis MacGregor based 
upon material from the Upper Llandeilo of the Berwyn Hills despite the fact that 
measurements (MacGregor 1961 : 196) indicate the type specimens at least to be very 
much wider. The significance of this difference cannot be ascertained until more is 
known about the variability of shell outline. MacGregor's description of 0. nantensis 
also suggests that the Shelve specimens have a greater number of ribs on the fold. 
However a photograph of the holotype {Ibid. pi. 20, fig. 18) shows 6 ribs on the 
fold in contradiction to an error in description which speaks of '3-5 ribs on the dorsal 
fold' {Ibid. : 197) ; it is possible that modal rib densities for the Shelve and Berwyn 
Oxoplecia do not greatly differ. 



Order STROPHOMENIDA Opik 1934 

Suborder STROPHOMENIDINA Opik 1934 

Superfamily PLECTAMBONITACEA Jones 1928 

Family LEPTESTIIDAE Opik 1933 

Subfamily LEPTESTIINAE Opik 1933 

PALAEOSTROPHOMENA Holtedahl 

Palaeostrophomena sp. 

(PI. 2i, figs. 7, io, 11, 13) 

Diagnosis. Semicircular Palaeostrophomena just over half as long as wide, 
ornamented by unequally developed parvicostellae numbering 8 to 10 per mm, 
5 mm antero-medially of the umbo, very fine fila and about 8 impersistently con- 
centric rugae ; ventral muscle scar bilobed divided by divergent vascula media. 

Description. Semicircular Palaeostrophomena, with a pedicle valve 53% as 
long as wide, acute cardinal angles and a very gently concavo-convex profile flatten- 
ing or slightly resupinate peripherally ; shell surface ornamented by very fine 
crowded fila, about 8 concentric rugae with a wavelength of about 0-2 mm strongly 
developed at acute angles to the hinge-line but becoming subdued anteriorly, and 
by parvicostellae numbering 8 to 10 per mm, 5 mm antero-medially of the umbo, 
and divided into narrow sectors by accentuated ribs, 10 or 11 of which arise early in 
the umbonal area ; ventral and shorter dorsal interareas apsacline and anacline 
respectively with a small pseudodeltidium and a supra-apical foramen, and a dorsal 
notothyrium filled by a median ridge representing the posterior surface of a plectam- 
bonitacean cardinal process. 

Ventral interior with relatively small simple teeth, vestigial dental plates and a 
bilobed ventral muscle scar extending forward for about one-fifth the length of the 
valve and about three-fifths as long as wide ; diductor scars extending beyond the 



ORDOVICIAN BRACHIOPODA 127 

median adductor impression divided by divergent vascula media. Dorsal interior 
unknown. 

Figured material. 

length width (mm) 
External and internal moulds of pedicle valve (BB 35415a, b) 8-0 14-0 

External and internal moulds of pedicle valve with adherent 
shell (BB 35451a, b) 10-5 

External moulds of shell (BB 35414a, b) 7*5 14-0 

Horizon and localities. Whittery Shales : BB 35415 from exposures half-way 
down path below Marrington Farm going southwards (Grid Ref. SO 272967) ; 
BB 35451 from the lower part of the succession in the west bank of the River Camlad, 
60 yds north-east of Marrington Farm (Grid Ref. SO 272970) ; BB 35414 from 
Whittery Quarry at the south end of Whittery Wood near Chirbury (Grid Ref. 
SO 275981). 

Discussion. Despite the apparent absence of accessory teeth in the pedicle valve 
and the lack of information about the dorsal interior, a few strophomenide moulds 
recovered from the Whittery Shales may be safely assigned to Palaeostrophomena in 
the wider interpretation of that genus (Williams in Whittington & Williams 1955 : 
415). In known characteristics, the specimens compare closely with P. magnified 
from the Derfel Limestone of N. Wales as well as certain Scottish species from the 
Ardmillan Series of Scotland (Williams 1962 : 158), so that specific identification is 
not possible until more material is available. 



Family EOCRAMATIIDAE nov. 

Diagnosis. Plectambonitaceans (?) with a complementary pseudodeltidium and 
chilidium, a small submesothyridid foramen, unsupported simple teeth, strong 
narrowly divergent chilidial plates (?) filling the notothyrium in place of a discrete 
platform and continuous with widely splayed cup-shaped socket ridges ; shell 
impunctate. 

Discussion. There are many features of the new genus Eocramatia that reflect 
the tendency of related major groups to show morphological convergence as they are 
traced back towards their common ancestry. The presence of a pseudodeltidium 
and chilidium indicates affinities with the Clitambonitidina, Strophomenidina and 
Billingsellacea. But the absence of dental plates and the structure of the cardinalia 
preclude assignment of the genus to the Clitambonitidina, and although the im- 
punctate condition of the shell and the morphology of the pedicle valve suggest that 
Eocramatia was derived from the billingsellaceans, many dorsal features contradict 
this interpretation. These include the absence of an orthide notothyrial platform, 
and the nature of the divergent plates filling the notothyrium and their relationship 
to the socket ridges. In all, then, it seems that it can be fairly confidently assigned 
to the Strophomenidina. 



128 SHELVE DISTRICT 

Although only plectambonitaceans are currently known to have been contem- 
poraneous with Eocramatia, representatives of the other two strophomenidine super- 
families, the Strophomenacea and Davidsoniacea, first occur in the Upper Llanvirn 
and basal Caradoc respectively and are stratigraphically eligible for consideration 
as related stocks. From a general morphological viewpoint the pseudopunctate 
strophomenaceans are unlikely to be related, but whether the new genus should be 
assigned to the plectambonitaceans or the davidsoniaceans mainly depends on the 
interpretation of the plates filling the notothyrium. These structures are like the 
chilidial plates of the leptellids and, in conjunction with the planar brachial valve 
and the absence of a notothyrial platform, suggest that the new genus is a plectam- 
bonitacean. On the other hand, the posterior surfaces of the plates appear to have 
been grooved in life and, therefore, may have functioned as supports for the dorsal 
diductor muscle bases which were, in contrast, attached to the notothyrial floor 
between the chilidial plates in plectambonitaceans. If this were so the plates are 
homologous with the bilobed cardinal process and, together with the impunctate 
shell condition and simple teeth, hint at a davidsoniacean relationship. The 
principal features refuting this interpretation are the attitude of the socket ridges, 
the absence of dental plates and the planar brachial valve ; and, since simple teeth 
without flanking accessory processes and an impunctate shell are known to have 
persisted among the most primitive plectambonitaceans such as the contemporaneous 
plectambonitids and taffiids, it seems more appropriate to assign Eocramatia to the 
Plectambonitacea at least until better preserved material is available to decide the 
issues in question. 

Having had so much difficulty in assigning Eocramatia to a superfamily it is not 
surprising to find that its familial connections within that taxon are at present elusive. 
In some respects Eocramatia is closest to the Taffiidae, but both the impunctate 
shell and the structure of the cardinalia militate against including it in that family 
and, in association with other features, call for the erection of a new monotypic 
family. 



Genus EOCRAMATIA nov. 

Name. An early brachiopod with a mixture (Gk. Kpdfia) of morphological features. 

Diagnosis. Subquadrate, slightly uniplicate plano-convex shells widest at the 
hinge-line, with a shallow median sulcus in the pedicle valve corresponding to a low 
median fold in the obscurely sulcate brachial valve ; ornamentation costellate by 
branching and intercalation tending to thicken between intersections with concentric 
lamellae ; ventral interarea high, curved apsacline, pseudodeltidium long, arched, 
foramen submesothyridid, dorsal interarea planar hypercline, chilidium small arch- 
ing over the dorsal ends of high, thick, narrowly divergent chilidial plates ; shell 
impunctate. 

Ventral interior with flat semi-oval teeth embedded in thick secondary shell, 
dental plates absent, ventral muscle field indistinct but apparently not extending 
much beyond delthyrial cavity ; mantle canal system obscure except for subparallel 



ORDOVICIAN BRACHIOPODA 



129 



vascula media and terminal branches but possibly associated as a pinnate pattern 
with low thin ridges radiating from vicinity of delthyrial cavity. 

Dorsal interior with flattened, slightly indented ventral ends of chilidial plates 
filling notothyrial cavity and continuous with slightly convex plate-like socket 
ridges disposed parallel to hinge-line ; notothyrial platform absent so that cardinalia 
overhang a shallow posteromedian depression passing anteriorly into low narrow 
median ridge ; adductor muscle field quadripartite bounded by outwardly curving 
raised ridges ; dorsal mantle canal pattern preserved only peripherally, possibly 
pinnate. 

Type species. Eocramatia dissimulata sp. nov. from the Hope Shales. 

Eocramatia dissimulata sp. nov. 
(PI. 21, figs. 9, 12, 14, 15 ; PI. 22, figs. 1-3, 5, 6 ; Text-fig. 10) 

Diagnosis. Subquadrate, uniplicate, plano-convex Eocramatia with the pedicle 
valve 74% as long as wide and 18% as deep as long and ornamented by costellae 
numbering 5 per mm, 5 mm antero-medially of the ventral umbo ; dorsal adductor 
scars limited to the posterior half of the valve. 

Description. Subquadrate, plano-convex Eocramatia with 3 pedicle valves 
averaging 74% as long as wide (range 68% to 81%) and 18% as deep as long (range 
17% to 19%) bearing a shallow median sulcus less than one-third as wide as the 
hinge-line and a brachial valve nearly three-fifths as long as wide bearing a vague 
median fold, cardinal margins acute ; costellate ornamentation fine, numbering 
5 per mm, 5 mm antero-medially of the umbones of 3 pedicle valves ; apsacline 
ventral interarea 19% as long as the pedicle valve (average for 4 valves, with a range 
of 15% to 22%) with the pseudodeltidium extending forward for about one-seventh 
the length of valve, hypercline dorsal interarea shorter. 



foramen 




chilidial plate C) 



socket ridge 
tooth 



muscle scar 




A B 

Fig. 10. Diagrammatic views of (A) the ventral interior and (B) the dorsal interior, 
with a posterior view of the brachial valve above, of Eocramatia. 



130 SHELVE DISTRICT 

Ventral interior with widely placed teeth extending laterally on either side of the 
delthyrium for about one-third the width of the pedicle valve, ventral subperipheral 
rim ill-defined. Dorsal interior with sockets elongated parallel to the hinge-line, 
dorsal adductor impressions limited to posterior half of the brachial valve. 
Type material. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 35488a, b) 6-8 9-0 

Paratypes External and internal moulds of brachial valve 

(BB 35489a, b) 5-2 8-o 

External and internal moulds of pedicle valve 

(BB 35490a, b) 5-5 7-5 

Incomplete external and internal moulds of 

conjoined valves (BB 35491a, b) - 10-5 

External mould of brachial valve (BB 35492) 5-0 - 

Horizon and locality. Hope Shales cropping out in Brithdir farmyard, 1 mile 
ENE of Old Church Stoke (Grid Ref. SO 301953). 

Discussion. No other species has yet been described that can profitably be 
compared with E. dissimulate/., and as the new species is represented by only a small 
number of moulds few statistics can be compiled from the sample to indicate the 
variability of diagnostic features. These are incorporated in the description. 



Family SOWERBYELLIDAE Opik 1930 
Subfamily SOWERBYELLINAE Opik 1930 

SOWERBYELLA Jones 1928 
Sowerbyella antiqua Jones emended A. W. 

(PL 22, figs. 4, 7-14 ; PI. 23, figs. 1, 3, 4) 

1928 Sowerbyella antiqua Jones : 419. 

Diagnosis. Semicircular Sowerbyella with a brachial valve 51% as long as wide, 
external surface very rarely ornamented by impersistent rugae finely parvicostellate 
with a modal count of 10 ribs per mm, 2 mm antero-medially of the dorsal umbo, 
infrequently segregated into sectors with a mean width of 0-57 mm ; ventral muscle 
scar extending forward for 32% of the length of the pedicle valve ; dorsal interior with 
a median septum usually flanked by 3 pairs of lateral septa, extending anteriorly for 
56% of the length of the brachial valve and commonly united into a raised platform 
at the 4 mm growth stage. 

Description. Semicircular Sowerbyella with acute cardinal angles, planar or 
very gently concave brachial valve 51% as long as wide, uniformly convex pedicle 
valve 27% as deep as long (range 14% to 34% for 11 specimens) ; ventral and dorsal 
interareas apsacline and anacline respectively with a supra-apical foramen and a 
small pseudodeltidium and a complementary chilidium ; ornamentation almost 



ORDOVICIAN BRACHIOPODA 131 

exclusively parvicostellate with counts of 9, 10, 11 and 12 ribs per mm, 2 mm antero- 
medially of the umbones of 6, 30, 19 and 9 brachial valves and differentiated into 
sectors with a mean width of 0-57 mm, 2 mm antero-medially of the umbo in 24 out 
of 64 valves ; rugae with a wavelength of 0-15 mm rarely developed sporadically in 
postero-lateral areas. 

Teeth small, dental plates obsolescent ; ventral muscle field strongly cordate in 
outline, 74% as long as wide and extending forward for 32% of the length of the 
pedicle valve ; adductor scars small, impressed as two hollows in the secondary 
shell of the umbo and divided by a fine median ridge ; diductor scars splayed 
anteriorly and divided by divergent vascula media of a lemniscate mantle canal 
system. 

Cardinalia consisting of a median cardinal process fused with chilidial plates and 
widely divergent socket plates extending anteriorly for 36% of their lateral spread 
and for 12% of the length of the brachial valve ; notothyrial platform poorly devel- 
oped ; lophophore and muscle supports consisting of a median septum almost 
invariably flanked by 3 pairs of lateral septa and extending anteriorly for 56% of the 
length of the brachial valve, commonly ankylosed to form an elevated cleft platform 
in adult valves more than 4 mm long. 

Figured material. 

External mould of brachial valve (BB 35524) 
Incomplete internal mould of brachial valve (BB 35525) 
External mould of brachial valve (BB 35526) 
Internal mould of pedicle valve (BB 35527) 
External mould of pedicle valve (BB 35528) 
Incomplete internal mould of brachial valve (BB 35529) 
Internal mould of brachial valve (BB 35530) 
Incomplete internal mould of brachial valve (BB 35531) 
Internal mould of brachial valve (BB 35532) 
Internal mould of pedicle valve (BB 35533) 
Incomplete internal mould of pedicle valve (BB 35534) 

Horizon and locality. Flags of the Ffairfach Group exposed on the western 
side of Coed Duon near Llangadog (Grid Ref. SN 709256). 

Sowerbyella cf. antiqua Jones 
(PL 23, figs. 2, 5-13) 

Diagnosis. Like Sowerbyella antiqua but with a ventral muscle scar extending 
forward for only 26% of the length of the pedicle valve. 

Description. Semicircular, concavo-convex Sowerbyella with acute cardinal 
angles, brachial valve 50% as long as wide, pedicle valve 15% as deep as long ; 
ventral and dorsal interareas apsacline and anacline respectively with a supra- 
apical foramen and a small pseudodeltidium and complementary chilidium ; orna- 
mentation exclusively parvicostellate with counts of 8, 9, 10, 11 and 12 ribs per mm, 



length 


width (mm) 


4'5 


9-2 


5-0 


io-o 


6-0 


8-5 


6-o 


- 


3-o 


- 


3'4 


7-0 


4-6 


- 


5-o 


io-o 


5-o 


9-0 


2'5 


- 



132 SHELVE DISTRICT 

2 mm antero-medially of the umbones of 3, 6, 10, 7 and 1 brachial valves respectively, 
and segregated into sectors by thickened costellae in only 2 out of 27 valves. 

Teeth small, dental plates obsolescent, cordate ventral muscle field 62% as long 
as wide and extending forward for 26% of the length of the pedicle valve ; muscle 
scars and mantle canal system differentiated as in S. antiqua s.s. 

Cardinalia consisting of a median cardinal process fused with chilidial plates and 
widely divergent socket plates extending anteriorly for 10% of the length of the 
brachial valve and for an average of 28% of their lateral spread in 6 valves (range 
2 5% to 35%) ; notothyrial platform poorly developed ; lophophore and muscle 
supports consisting of a median septum flanked by 3 pairs of lateral septa, rarely 
more or fewer, united by secondary shell into a cleft platform in 2 out of 13 specimens 
over 3-5 mm long. 

Figured material. 

External mould of brachial valve (BB 35535) 

Incomplete external mould of brachial valve (BB 35536) 

Internal mould of pedicle valve (BB 35537a) 

External mould of pedicle valve (BB 35537b) 

Internal mould of brachial valve (BB 35538) 

Internal mould of pedicle valve (BB 35539) 

Incomplete internal mould of pedicle valve (BB 35540) 

Internal and incomplete external moulds of brachial valve 

(BB 3554ia, b) 

Incomplete internal mould of brachial valve (BB 35542) 

Horizon and locality. Flags transitional from the Meadowtown to Betton 
Beds exposed in a ploughed field 180 yds south-west of a wall in side of lane from 
Meadowtown to Castle Ring (Grid Ref. SJ 310009). 

Sowerbyella multiseptata sp. nov. 

(PI. 23, figs. 14-19 : PL 24, figs. 1, 3) 

Diagnosis. Semicircular Sowerbyella with a brachial valve 51% as long as wide, 
external surface very rarely ornamented by impersistent rugae, finely parvicostellate 
with a modal count of 10 ribs per mm, 2 mm antero-medially of the dorsal umbo ; 
ventral muscle scar extending forward for 32% of the length of the pedicle valve ; 
dorsal interior with a late-developing median septum usually flanked by 3 pairs of 
lateral septa extending anteriorly for 73% of the length of the brachial valve. 

Description. Semicircular Sowerbyella with acute, rarely orthogonal, cardinal 
angles ; brachial valve 51% as long as wide uniformly gently concave, occasionally 
with a narrow median fold extending from a protegulum 0*2 mm long to die out 
anteriorly ; pedicle valve convex, rarely subcarinate, with a mean depth of 17% 
relative to the length of 4 valves (range 16% to 19%) ; ventral and dorsal interareas 
apsacline and anacline respectively with a supra-apical foramen and a small pseudo- 
deltidium and complementary chilidium ; ornamentation almost exclusively finely 
parvicostellate with counts of 8, 9, 10 and 11 ribs per mm, 2 mm antero-medially of 



length 


width (mm) 


2-5 


5-5 


2-3 


- 


3-o 


6-o 


2-8 


5-8 


2-5 


5-2 


3-o 


5-5 


4-3 


— 


4-3 


8-o 


2-5 


- 



ORDOVICIAN BRACHIOPODA 133 

the umbones of 1, 1, 7 and 4 brachial valves respectively ; impersistent postero- 
lateral rugae very rare with a wavelength of 0-25 mm.. 

Teeth small, dental plates obsolescent, cordate ventral muscle field 71% as long 
as wide and extending forward for 32% of the length of the pedicle valve ; adductor 
scars small, impressed in 2 hollows in the umbonal secondary shell and divided by a 
fine median ridge, splayed diductor scars divided by divergent vascula media. 

Cardinalia consisting of a median cardinal process fused with chilidial plates and 
widely divergent socket plates extending anteriorly for an average of 30% of their 
lateral spread and for 10% of the length of 3 brachial valves ; notothyrial platform 
poorly developed ; lophophore and muscle supports extending anteriorly for 73% 
of the length of the brachial valve and consisting of a relatively short median septum 
developing later than a pair of strong submedial septa normally with 2 additional 
pairs of septa occurring lateral of the median triad, septa becoming united into a 
cleft platform in 2 out of 10 specimens over 2-5 mm long. 

Type material. 

length width (mm) 

Holotype Internal mould of brachial valve (BB 35544) 2-8 5-3 

Paratypes External and internal moulds of pedicle valve 

(BB 35545a, b) 2-6 4-8 

Internal and incomplete external moulds of 

pedicle valve (BB 35546a, b) 3-0 6-0 

Internal and incomplete external moulds of 

brachial valve (BB 35547a, b) 3-5 6-5 

External and internal moulds of pedicle valve 

(BB 35548a, b) 5-3 9-5 

Incomplete internal mould of brachial valve 

(BB 35549) 1-5 

Horizon and localities. Spy Wood Grit : BB 35547 from exposures on top of 
ridge 1440 yds NNE of Rorrington bench mark 599 (Grid Ref. SJ 303018) ; all other 
specimens from exposures 1100 yds NNE of Rorrington (Grid Ref. SJ 303015). 

Sowerbyella cf. sericea permixta Williams 

(PI. 24, figs. 2, 4-7, 10) 

1963 Sowerbyella sericea (J. de C. Sowerby) permixta Williams : 434. 

Diagnosis. Semicircular Sowerbyella with a brachial valve 50% as long as wide, 
external postero-lateral surfaces infrequently ornamented by up to 7 pairs of im- 
persistent rugae with a wavelength of 0-3 mm and by fine parvicostellae with modal 
counts of 10 per mm, 2 mm antero-medially of the dorsal umbo and commonly 
differentiated into sectors with a mean width of 0-67 mm ; ventral muscle scar 
extending forward for 42% of the length of the pedicle valve ; dorsal interior with a 
median septum occurring rarely and developing late, submedial septa strong, usually 
flanked by a low platform with or without the differentiation of 2 pairs of lateral 
septa and extending anteriorly for 66% of the length of the brachial valve. 



134 SHELVE DISTRICT 

Description. Semicircular Sowerbyella with acute cardinal angles becoming 
roundly obtuse in late growth stages, planar to gently concave brachial valve 50% 
as long as wide, pedicle valve 19% as deep as long, evenly convex to subcarinate in 
transverse profile ; ventral and dorsal interareas apsacline and anacline respectively 
with a supra-apical foramen and a small pseudodeltidium and complementary 
chilidium ; radial ornamentation normally unequally parvicostellate (in 25 out of 
32 brachial valves) with rib counts of 8, 9, 10, 11 and 12 per mm, 2 mm antero- 
medially of the umbones of 2, 7, 15, 7 and 1 brachial valves respectively, with a mean 
width of sectors delineated by thickened costellae of 0-67 mm ; up to 7 pairs of rugae 
with a wavelength of 0-3 mm usually poorly and impersistently developed in the 
postero-lateral areas at acute angles to the hinge-line. 

Teeth small, dental plates obsolescent, cordate ventral muscle field 71% as long 
as wide and extending anteriorly for 42% of the length of the pedicle valve ; adductor 
scars small, deeply inserted in a pair of hollows in the umbonal secondary shell and 
divided by a fine median ridge ; diductor scars splayed anteriorly and divided by 
vascula media forming part of a lemniscate mantle canal system. 

Cardinalia consisting of a median cardinal process fused with chilidial plates and 
widely divergent socket plates extending anteriorly for 33% of their lateral spread 
and for 13% of the length of the brachial valve ; notothyrial platform poorly devel- 
oped ; lophophore and muscle supports normally consisting of 3 pairs of septa with 
a strong submedial pair extending anteriorly for 66% of the length of the brachial 
valve making up a low cleft platform ; a short median septum was developed in only 
3 out of 20 valves less than 4 mm long and in 9 out of 24 valves more than 4 mm long. 

Figured material. 

length width (mm) 

External and internal moulds of pedicle valve (BB 35551a, b) 7-0 13-0 
External and internal moulds of brachial valve (BB 35552a, b) 3-8 7-5 

External and internal moulds of brachial valve (BB 35553a, b) 4-2 8-2 

External and internal moulds of brachial valve (BB 35554a, b) 47 9-0 

Horizon and locality. Aldress Shales exposed in the bank of Ox Wood Dingle 
at the south-west corner of Ox Wood a few yards north of the Rorrington-Wotherton 
road (Grid Ref. SJ 290007). 



Sowerbyella cf. sericea (J. de C. Sowerby) 
(PL 24, figs. 11, 13, 14, 16) 

1839 Leptaena sericea J. de C. Sowerby : 636. 

1928 Sowerbyella sericea (J. de C. Sowerby) Jones : 414. 

1963 Sowerbyella sericea (J. de C. Sowerby) Williams : 430. 

Diagnosis. Semicircular Sowerbyella with a brachial valve 50% as long as wide 
and a pedicle valve 26% as deep as long, ornamented by up to 9 pairs of impersistent 
rugae with a wavelength of 0*3 mm, and unequally developed parvicostellae with a 
modal count of 9 per mm, 2 mm antero-medially of the dorsal umbo ; ventral muscle 
scar extending forward for 30% of the length of the pedicle valve ; dorsal interior 



ORDOVICIAN BRACHIOPODA 135 

with strong submedian septa extending anteriorly for 69% of the length of the 
brachial valve. 

Description. Semicircular Sowerbyella with obtusely rounded cardinal angles in 
adult shells, gently concave brachial valve 50% as long as wide (with a range of 47% 
to 52% for 5 valves), pedicle valve with a mean depth relative to length of 26% for 
6 valves (range 18% to 31%) with steeply sloping sides ; ventral and dorsal interareas 
apsacline and anacline respectively with a supra-apical foramen and a small pseudo- 
deltidium and complementary chilidium ; radial ornamentation unequally parvi- 
costellate with counts of 7, 8, 9, 10 and 11 ribs per mm, 2 mm antero-medially of the 
umbones of 1, 2, 4, 3 and 1 brachial valves respectively, with sectors delineated by 
thickened costellae having a mean width of 0-93 mm (for 8 estimates with a range 
of 0-7 to 1-2 mm) ; up to 9 pairs of rugae with a modal wavelength of 0-3 mm occur 
in the postero-lateral areas at acute angles to the hinge-line. 

Teeth small, dental plates obsolescent, cordate ventral muscle field 65% as long 
as wide and extending anteriorly for an average of 30% of the length of 6 pedicle 
valves (range 24% to 38%) ; adductor scars deeply inserted in a pair of hollows in the 
umbonal secondary shell and divided by a fine median ridge ; diductor scars splayed 
anteriorly and divided by vascula media forming part of a lemniscate mantle canal 
system. 

Cardinalia consisting of a median cardinal process fused with chilidial plates and 
widely divergent socket plates extending anteriorly for 35% of their lateral spread 
and for an average of 14% of the length of 3 brachial valves ; notothyrial platform 
poorly developed ; lophophore and muscle supports normally consisting of a semi- 
circular platform with strong septa bounding a median cleft extending anteriorly 
for 69% of the length of brachial valves (range 60% to 75%). 

Figured material. 

length width (mm) 
Internal mould of pedicle valve (BB 35555) 87 14-0 

External and internal moulds of brachial valve (BB 35556a, b) 6-3 12-0 
Pedicle valve (BB 35557) 9' 13*5 

Horizon and localities. Whittery Shales : BB 35555 from exposures in the 
lower part of the west bank of the River Camlad, 60 yds north-east of Marrington 
Farm (Grid Ref. SO 272970) ; BB 35556 from Whittery Quarry at the south end of 
Whittery Wood, near Chirbury (Grid Ref. SO 275981) ; BB 35557 from exposures 
in the stream at the north end of Spring Coppice 865 yds south-east of Hockleton 
Bridge (Grid Ref. SO 279997). 

Discussion. Species of Sowerbyella found in the Shelve succession belong to two 
distinct supra-specific groups which may eventually prove to be worth recognizing 
taxonomically in the manner proposed by Roomusoks (1959 : 43), although S. 
antiqua and 5. multiseptata are not characterized by a differentiated radial ornamen- 
tation like that found in the otherwise similar type species of his subgenus Viruella. 
Those taken from the Meadowtown Beds and Spy Wood Grit are characterized by a 
finely parvicostellate ornamentation only infrequently differentiated into sectors 
delineated by thickened ribs, and by the rare development of rugae, acutely disposed 



136 SHELVE DISTRICT 

to the hinge-line, in the posterolateral areas. The dorsal interiors of mature valves 
are further distinguished by the almost invariable presence of a median septum 
flanked by 3 pairs of septa in various stages of coalescence. The occurrence of a 
median septum in these stratigraphically older Sowerbyella is especially significant 
although the timing of its appearance in relation to the submedial septa was different. 
In the Meadowtown sample, the median septum is at least as strongly developed as 
the submedial pair even in the smallest valve (1-5 mm long), which suggests that the 
septa were secreted more or less simultaneously presumably after the transformation 
of the trocholophous lophophore into a schizolophe. In the Spy Wood sample, on 
the other hand, the median septum is shorter than the submedial pair, and in 3 moulds 
less than 2 mm long it was either absent or weak indicating that it developed later 
than the submedial septa. 

At this juncture it can be stated that the Meadowtown specimens are indistinguish- 
able from a sample of S. antiqua Jones from Ffairfach flags of the Llandeilo area, 
not only in the early development of the median septum but in every other mor- 
phological feature except for the greater length of the ventral muscle scar in 5. 
antiqua s.s. (p < o-ooi). Although this difference is statistically significant, it 
seems reasonable to compare the Meadowtown material with the Welsh species. 

The Spy Wood specimens differ significantly from both samples of S. antiqua not 
only in the late development of the median septum but also in the relatively faster 
anterior spread of the septa (0-02 > p > o-oi), and in the slower lateral growth of 
the brachial valve relative to its increase in length (o-oi > p > o-ooi). These 
differences are sufficient to warrant the recognition of a new species, S. multiseptata, 
based on the Spy Wood sample. 

The second distinctive group of Sowerbyella includes specimens recovered from the 
younger Aldress Shales and the Whittery and Hagley Shales. Only the Aldress 
sample is big enough to indicate the range of septa development. It does, however, 
demonstrate a further significant reduction in the development of a median septum 
compared with even its relatively late appearance in S. multiseptata. Only a pair of 
strong submedial septa occurred in 4 moulds less than 2-5 mm long ; and a short septum 
had developed in only 3 out of 24 larger valves. No immature specimens of Sower- 
byella were recovered from the Whittery and Hagley Shales, but traces of a septum 
were not seen in any of the 4 brachial valves examined which were between 3 and 
6-6 mm long. 

Comparisons of the Aldress and Whittery samples with described Caradocian 
Sowerbyella indicate that they are closely comparable with S. sericea (J. de C. 
Sowerby). The Aldress specimens in particular could be well tested statistically in 
relation to recently described species (Williams 1963 : 430-447). It was found that 
they differed from S. sericea permixta Williams only in the relatively wider spread 
of the ventral muscle scar (p < o-ooi). The Whittery and Hagley specimens, 
however, are more like 5. sericea s.s. in the coarser parvicostellate ribbing, the wave- 
length, number and more regular occurrence of the postero-lateral rugae, and the 
greater relative depth of the pedicle valve. Provisionally, therefore, the strati- 
graphically youngest Sowerbyella are referred to S. sericea s.s. and the Aldress forms 
to the subspecies permixta. 



ORDOVICIAN BRACHIOPODA 



137 



Table 98 

Statistics of length (1) and maximum width (w) of n brachial valves of S. antiqua Jones from the 

Ffairfach Group (A), S. cf. antiqua from the Meadowtown Beds (B), S. multiseptata sp. nov. from 

the Spy Wood Grit (C) and S. cf . sericea permixta Williams from the Aldress Shales (D) 





A 


B 


C 


D 


n 


44 


3i 


10 


20 


1 mm 


377 


2-66 


2-27 


3-94 


(var 1) 


(o-8 5 i) 


(0-508) 


(o-333) 


(3-096) 


w mm 


7'45 


5-36 


4-42 


7-83 


(var w) 


(3-°39) 


(2-112) 


(0-971) 


(12-016) 


r 


0-956 


0-989 


0-987 


o-995 


logel 


1-2988 


0-9427 


0-7885 


1-2817 


(var logel) 


(0-0579) 


(0-0695) 


(0-0626) 


(0-1815) 


logeW 


1-9816 


I-6437 


1-462 


1-9685 


(var log e w) 


(0-0532) 


(0-0708) 


(0-0484) 


(0-1789) 


r e 


0-964 


o-995 


0-992 


0996 


a 


0-9582 


1-0094 


0-8795 


0-9928 


(var a) 


(0-00153) 


(0-00036) 


(0-00161) 


(0-00046) 



Table 99 

Statistics of length (1) and maximum depth (th) of n pedicle valves of 5. cf. antiqua from the 
Meadowtown Beds (B) and S. cf. sericea permixta Williams from the Aldress Shales (D) 

(see Table 98) 

B D 

n 10 9 

3-28 (0-491) 5-n (3-I94) 

0-49 (0-023) 0-97 (0-232) 

0-883 0795 



1 mm (var 1) 
th mm (var th) 

r 
a (var a) 



0-2175 (0-0013) 0-2698 (0-00381) 



Table ioo 

Statistics of length of pedicle valve (1) and length of ventral muscle scar (sc) in n examples of S. 
antiqua Jones from the Ffairfach Group (A), 5. cf. antiqua from the Meadowtown Beds (B), S. 
multiseptata sp. nov. from the Spy Wood Grit (C) and S. cf. sericea permixta Williams from the 

Aldress Shales (D) 





A 


B 


C 


D 


n 


37 


20 


15 


21 


1 mm 


4-71 


3-24 


3H5 


5-2 


(var 1) 


(0-831) 


(0-869) 


(0-941) 


(3-ioi) 


sc mm 


1-5 


0-85 


1-09 


2-2 


(var sc) 


(0-157) 


(0-119) 


(0-146) 


(I-I09) 


r 


0-885 


o-947 


0-967 


o-953 


logel 


1-532 


I-I359 


I-20I 


1-5954 


(var logel) 


(0-0367) 


(0-0794) 


(0-076) 


(0-1083) 


log e SC 


0-3697 


-0-239 


0-0316 


0-6878 


(var logeSc) 


(0-0675) 


(o-i53) 


(0-H53) 


(0-2055) 


r e 


0-898 


o-954 


0-969 


0-9611 


a 


1-3567 


1-3878 


I-23I7 


1-3776 


(var a) 


(0-01014) 


(0-00959) 


(0-007I2) 


(0-00761) 



138 shelve district 

Table ioi 

Statistics of length (1) and width (w) of ventral muscle scar in n pedicle valves of S. antiqua Jones 

from the Ffairfach Group (A), S. cf. antiqua from the Meadowtown Beds (B), S. multiseptata 

sp. nov. from the Spy Wood Grit (C), and S. cf. sericea permixta Williams from the Aldress 

Shales (D) 





A 


B 


c 


D 


n 


33 


15 


7 


18 


1 mm 


I-5I 


0-83 


1-16 


2-32 


(var 1) 


(o-i54) 


(0-081) 


(0-209) 


(1-054) 


w mm 


2-04 


i-33 


1-63 


3-28 


(var w) 


(0-216) 


(0-167) 


(o-339) 


(i-58i) 


r 


0-869 


0-974 


o-975 


0-961 


logel 


0-3767 


-0-2373 


0-0734 


0-7534 


(var logel) 


(0-0656) 


(o-iioi) 


(o-i45i) 


(0-1783) 


log e W 


0-6873 


0-2429 


0-4276 


1-1204 


(var log e w) 


(0-0506) 


(0-0895) 


(0-1202) 


(0-1368) 


r e 


0-874 


o-975 


0-977 


0-965 


a 


0-8782 


0-9018 


0-91 


0-8758 


(var a) 


(0-00585) 


(0-00306) 

Table 102 


(0-00756) 


(0-00325) 



Statistics of length (1) and maximum anterior extension of socket ridges (lc) in n brachial valves 

of 5. antiqua Jones from the Ffairfach Group (A), S. cf. antiqua from the Meadowtown Beds (B) 

and S. cf. sericea permixta Williams from the Aldress Shales (D) 





A 


B 


D 


n 


40 


20 


32 


1 mm 


4-26 


2-97 


4-13 


(var 1) 


(0-716) 


(0-684) 


(1-768) 


lc mm 


o-5 


0-28 


0-52 


(var lc) 


(0-016) 


(0-012) 


(0-041) 


r 


0-918 


o-945 


0-918 


logel 


1-4305 


1-0531 


1-3685 


(var logel) 


(0-0386) 


(0-0742) 


(0-0987) 


logelc 


-0-7301 


-I-3233 


-0-7273 


(var logelc) 


(0-0639) 


(0-1363) 


(0-1419) 


r e 


0-924 


o-954 


0-924 


a 


1-2867 


1-3554 


1-1991 


(var a) 


(0-00633) 


(0-00923) 


(0-00704) 



Table 103 

Statistics ofj maximum anterior (1) and maximum lateral (w) extensions of socket ridges in n 
brachial valves of S. antiqua Jones from the Ffairfach Group (A) and S. cf. sericea permixta 

Williams from the Aldress Shales (D) 





A 


D 


n 

1 mm (var 1) 
w mm (var w) 

r 


28 
0-52 (0-056) 
1-47 (0-107) 
0-678 


21 
o-54 (0-036) 
1-59 (0-209) 
0-826 


logel (var logel) 


-0-7365 (0-1844) 


-0-6789 (0-1186) 


log e w (var log c w) 

r e 
a (var a) 


o-3595 (0-0483) 

0696 

0-5117 (0-00519) 


0-4274 (0-0789) 

0-835 

0-8159 (0-0106) 



ordovician brachiopoda 139 

Table 104 

Statistics of length (1) and maximum anterior extension of septa (Is) in n brachial valves of 
5. antiqua Jones from the Ffairfach Group (A), S. cf. antiqua from the Meadowtown Beds (B), 
S. multiseptata sp. nov. from the Spy Wood Grit (C) and S. cf. sericea permixta Williams from 

the Aldress Shales (D) 





A 


B 


C 


D 


n 


54 


35 


13 


43 


1 mm 


4 -2 


2-98 


273 


4-07 


(var 1) 


(o-599) 


(o-59i) 


(0-424) 


(1-876) 


Is mm 


2-34 


i-8i 


1-98 


2-68 


(var Is) 


(o-i95) 


(0-2) 


(0-305) 


(o-86i) 


r 


0-926 


o-954 


0-915 


0-969 



a 0-5697 0-5817 0-8478 0-6774 

(var a) (0-00088) (0-00092) (0-01065) (0-00068) 



Subfamily AEGIROMENINAE Havlicek 1961 

SERICOIDEA Lindstrom 1953 

Sericoidea cf. abdita Williams 
(PL 24, figs. 8, 9, 12, 15, 17, 18 ; PL 28, fig. 16) 

1955 Sericoidea abdita Williams : 418. 

Diagnosis. Semicircular Sericoidea with a brachial valve 54% as long as wide 
and ornamented by unequally developed parvicostellae commonly numbering 11 or 
12 per mm, 2 mm antero-medially of the dorsal umbo ; median and submedial septa 
in adult brachial valves extending forwards for about two-thirds the valve length 
bounded anteriorly by an arc of septules. 

Description. Semicircular, concavo-convex Sericoidea with a concave brachial 
valve averaging 54% as long as wide in 6 valves (range 50% to 64%) and an evenly 
convex pedicle valve with a mean depth for 3 valves of 11-3% (range 8-7% to 13-6%) ; 
ventral and dorsal interareas apsacline and anacline, supra-apical foramen open in 
adult pedicle valves ; protegulum about 0-2 mm long ; radial ornamentation un- 
equally parvicostellate with rib frequencies of 11, 12 and 13 per mm, 2 mm antero- 
medially of the umbones of 3, 3 and 1 brachial valves, divided into sectors about 
0-5 mm wide by accentuated costae and costellae ; impersistent rugae with a wave- 
length of about o-i mm occur postero-laterally at acute angles to the hinge-line. 

Ventral interior with unsupported oblique teeth ; subquadrate ventral muscle 
scar about as wide as long and extending forward for 28% of the length of the pedicle 
valve, bounded by low subparallel ridges trailing anteriorly from the tooth bases 
and divided postero-medially by a thin median ridge. 

Dorsal interior with widely splayed socket ridges extending anteriorly for about 
8% of the length of the valve ; lophophore supports of adult valves consisting of a 
median septum flanked by a pair of submedial septa extending forward for 62-8% 
of the length of 6 brachial valves (range 54% to 80%) and a peripheral arc of about 
8 septules. 



i 4 o SHELVE DISTRICT 

Figured material. 

length width (mm) 
External and incomplete internal moulds of brachial valve 

(BB 35484a, b) 2-5 3-9 

Internal mould of brachial valve (BB 35485) 2-2 4-1 

External and internal moulds of brachial valve (BB 35486a, b) 1-3 2-6 

External and internal moulds of pedicle valve (BB 35487a, b) 2-2 - 

Horizon and locality. Hagley Shales : interbedded tuff exposed in north end 
of coppice 300 yds north-east of Calcot, Church Stoke (Grid Ref. SO 275961). 

Discussion. The only Sericoidea found in the Shelve area occur in fine-grained 
tuffs associated with the Hagley Shales and Volcanics. Specimens are not common 
but moulds of the brachial valves recovered show that the sample is closely related 
to S. abdita Williams from the Derfel Limestone of the Arenig area (Williams in 
Williams & Whittington 1955 : 418 ; Williams 1962 : 188). In the smallest brachial 
valves up to i-6 mm long the median septum and a peripheral arc of sharp tubercles 
were strongly developed but the submedial septa only incipiently so. Further growth 
included an acceleration in the secretion of the submedial septa to amalgamate with 
the median septum and an accentuation of the arc of tubercles into discrete septules. 
Both S. abdita from Wales and S. aff. abdita from the Balclatchie Group of Girvan 
were characterized by the development of a similarly differentiated and proportioned 
lophophore platform. Indeed the only differences between the Shelve Sericoidea 
and those from Wales and Scotland is that the ribbing of the former is slightly coarser 
and the brachial valve relatively longer, although on available evidence these 
differences are not significant. 

Superfamily STROPHOMENACEA King 1846 

Family STROPHOMENIDAE King 1846 

Subfamily FURCITELLINAE Williams 1965 

FURCITELLA Cooper 1956 

Furcitella sp. 

(PL 25, figs. 4, 5) 

The external and internal moulds of a strophomenid brachial valve (BB 35422a, b), 
from the Whittery Shales exposed in Whittery Quarry at the south end of Whittery 
Wood near Chirbury (Grid Ref. SO 275981), have been provisionally identified as 
Furcitella. The valve which was 7-5 mm long was almost semicircular in outline 
with slightly obtuse cardinal angles. The valve was also resupinate, being flat up 
to the 2.5 mm growth stage but thereafter becoming convex to attain an overall 
depth of 10% of the length. The external surface was ornamented by fine, equally 
developed costellae numbering 8 per mm, 5 mm antero-medially of the umbo and 
crossed by fine fila with frequencies of up to 18 per mm. The surface also undulated 
as impersistent rugae with wavelengths of about 0-5 mm. The internal features are 
poorly preserved but the cardinalia and muscle boundaries can be identified as 



ORDOVICIAN BRACHIOPODA 141 

including a bilobed cardinal process, widely divergent socket ridges, a notothyrial 
platform extending forward for less than one-third the length of the valve into a 
median ridge possibly bifurcating anteriorly, and a pair of divergent transmuscle 
septa. 

In respect of its ornamentation, profile and internal features, the brachial valve is 
not only identifiable as Furcitella but also closely comparable with the type species 
F. plicata from the Oranda Formation of Virginia (Cooper 1956 : 877), although on 
the basis of impressions of only one valve specific identification would not be 
reliable. 

MURINELLA Cooper 1956 

Murinella sp. 

(PI. 25, figs. 2, 3) 

The incomplete external and internal moulds of a pedicle valve (BB 35421a, b), 
collected from the Meadowtown Beds exposed along the lane to Lower Ridge from 
bench mark 754, Little Weston (Grid Ref. SO 293984), are best identified as the 
strophomenid Murinella. The valve was 6-5 mm long and, when complete, must 
have been semi-elliptical in outline with slightly acute cardinal angles, and evenly 
convex with only a faintly developed median carina, giving a maximum depth 
relative to length of about 10%. The ornamentation was finely costellate with 6 
ribs per mm at 5 mm antero-medially of the umbo, crossed by fine fila and coarser 
sporadically occurring growth lines. A pedicle foramen truncated the beak and the 
relatively small teeth protruding dorsally of the apsacline interarea were supported 
by short, widely divergent dental plates. The muscle scar was transversely elliptical, 
extending forward for about one-third the length of the valve and about two-thirds 
as long as wide. The various components of the muscle field are not identifiable 
but there is a conspicuous median septum culminating in a point at the anterior 
boundary of the scar and trailing forward beyond the boundary for a short distance 
as a low ridge. 

The nature of the radial ornamentation, the even convexity of the valve and 
especially the presence of a median septum in the muscle scar differentiate the valve 
as belonging to Murinella. Indeed except for the greater extension of the median 
septum beyond the anterior boundary of the muscle scar the valve compares closely 
with the type species for the genus, M. partita from the Bromide Formation of 
Oklahoma (Cooper 1956 : 848) : however, more material may demonstrate that the 
Shropshire form is quite distinct. 

Subfamily RAFINESQUININAE Schuchert 1893 

RAFINESQUINA Hall & Clarke 1892 

Rafinesquina delicata sp. nov. 

(PL 25, figs. 6-13) 

Diagnosis. Piano- to gently concavo-convex Rafinesquina with a pedicle valve 
8% as deep as long and a brachial valve 79% as long as wide, radial ornamentation 



142 SHELVE DISTRICT 

unequally parvicostellate with 7 ribs per mm, 10 mm antero-medially of the dorsal 
umbo ; ventral muscle scar feebly impressed subcircular, as long as wide and 
extending anteriorly for 28% of the length of the pedicle valve ; cardinalia delicately 
developed, notothyrial platform vestigial except in late growth stages. 

Description. Piano- to gently concavo-convex, transversely semi-elliptical 
Rafinesquina with cardinal angles normally orthogonal but also obtuse or acute 
especially the latter in young shells ; pedicle valve evenly convex, 8% as deep as 
long (range 7% to 10% for 6 valves), brachial valve flat or slightly concave, rarely 
with a dorsally deflected margin, 79% as long as wide with a strongly convex trilobed 
protegulum 07 mm long ; ventral interarea apsacline with a large supra-apical 
foramen and a small pseudodeltidium, dorsal interarea narrow anacline with a large 
convex, medially indented chilidium ; radial ornamentation unequally parvicostel- 
late, rarely with a thickened median rib, with counts of 6, 7, 8 and 9 ribs per 
mm, 10 mm antero-medially of the umbones of 10, 13, 3 and 1 brachial valves 
respectively. 

Ellipsoidal teeth obliquely disposed parallel with the widely divergent dental 
plates reaching anteriorly for 9% of the length of the pedicle valve ; ventral muscle 
field subcircular, 99% as wide as long and extending forward for 28% of the length 
of the valve, usually faintly developed and consisting of a broad adductor scar 
impressed on either side of a median ridge and a pair of flanking semicircular 
diductor scars ; internal surface rarely coarsely tuberculate, mantle canal system 
unknown. 

Cardinal process lobes delicate, plate-like, slightly splayed and ankylosed to widely 
divergent thin socket ridges extending anteriorly for 9% of the length of the brachial 
valve and 31% of their lateral extension ; concave surfaces of narrow sockets 
normally tuberculate in adult valves ; notothyrial platform poorly developed 
except as a low anchor-shaped feature in late adult growth stages so that the car- 
dinalia usually bound slight postero-medial hollows in younger valves ; median 
ridge and transmuscle septa rarely developed, extending forward for about two-fifths 
the length of the valve ; only the posterior pair of reniform adductor scars known, 
extending anteriorly for 30% of the length of the valve ; subperipheral rim, up to 
2 mm wide, rarely developed. 

Type material. 

Holotype External and internal moulds of brachial valve 

(BB 35558a, b) 
Paratypes External and internal moulds of pedicle valve 

(BB 35559a, b) 

External and internal moulds of pedicle valve 

(BB 35560a, b) 

External and internal moulds of pedicle valve 

(BB 35561a, b) 

Incomplete internal mould of brachial valve 

(BB 35562) 



length 


width (mm 


12-5 


17-0 


13-0 


14-5 


14-0 


- 


n-5 


12-0 



ORDOVICIAN BRACHIOPODA 143 

length width (mm) 
External and internal moulds of brachial valve 
(BB 35563a, b) io-o n-5 

External and internal moulds of brachial valve 
(BB 35564a, b) 7-5 9-0 

Horizon and localities. Meadowtown Beds : BB 35558 and BB 35564 from 
exposures 100 yds south-east of Minicop Farm (Grid Ref. SJ 315018) ; all other 
specimens from Quinton's Quarry in a field 200 yds north-east of Meadowtown 
Chapel (Grid Ref. SJ 312013). 

Discussion. The Rafinesquina found sporadically in the Betton Beds and 
occurring abundantly in the Meadowtown Beds (Tables 105-110) is especially 
characterized by the delicacy of its cardinalia, the weakness of its notothyrial plat- 
form and the smallness of its subcircular ventral muscle scar. Indeed if it were not 
for the presence of strongly developed dental plates, the species would have been 
more appropriately classified as Platymena (Cooper 1956 : 879). The nature of these 
features also serves to distinguish the new species from its closest relative, the Upper 
Llandeilo R. simplex (MacGregor 1961 : 205) which, in particular, has a better defined 
notothyrial platform and significantly larger ventral muscle scar. In the absence 
of statistical data from MacGregor' s account of R. simplex, no other difference may 
be categorically identified at present, although the brachial valve of the Welsh 
species may prove to be relatively wider than that of R. delicata. 



Rafinesquina sp. 

(PL 25, fig. 1 ; PL 26, fig. 1) 

The only specimens from the Whittery Shales referrable to Rafinesquina are 
incomplete complementary external and internal moulds of a brachial valve 
(BB 35496a, b) recovered from Whittery Quarry at the south end of Whittery Wood, 
near Chirbury (Grid Ref. SO 275981). The valve was gently concave with a promi- 
nent protegulum about 0-5 mm long and an unequally parvicostellate radial 
ornamentation with a density of 10 ribs per mm antero-medially of the umbo. 
Impersistent rugae up to 2 mm in wavelength were developed in arcs concentric to 
the umbo and immediately distinguish the specimen from the Meadowtown Rafines- 
quina. Internally the cardinal process lobes were ridge-like and slightly divergent 
and capped postero-medially by a strong chilidium. The socket ridges were narrow 
and low and the anchor-shaped notothyrial platform weakly developed with only 
one pair of transmuscle septa evident. 

A number of strophomenids have been recorded from the Caradocian rocks of 
England and Wales as Rafinesquina expansa (Sowerby) but the species requires 
revision because, as currently understood (Davidson 1871 : 312), it includes specimens 
that are not even congeneric ; and until more material is available no comparisons 
are trustworthy. 



i 4 4 SHELVE DISTRICT 

Table 105 

Statistics of length (1) and maximum width (w) of 37 brachial valves of Rafinesquina delicata 

sp. nov. 

1 mm (var 1) 6-92 (17715) 

w mm (var w) 875 (25-535) 

r 0-994 

logel (var logel) 17778 (0-3143) 

log e w (var log e w) 2-0247 (0-2880) 

r e 0-996 

a (var a) 0-9572 (0-00022) 

Table 106 

Statistics of length (1) and maximum anterior extension of socket ridges (lc) in 18 brachial valves 

of Rafinesquina delicata sp. nov. 

1 mm (var 1) 10-64 ( IO '557) 

lc mm (var lc) 0-98 (0-15) 

r 0-937 

a (var a) 0-095 (0-00006) 

Table 107 

Statistics of maximum anterior (1) and maximum lateral (w) extensions of socket ridges in 10 
brachial valves of Rafinesquina delicata sp. nov. 

1 mm (var 1) o-86 (0-198) 

w mm (var w) 2-61 (0-934) 

r 0-928 

logel (var logel ) -0-2694 (°' 2 374) 

log e w (var log e w) 0-8951 (0-1283) 

r e 0-937 

a (var a) 0-7352 (0-00819) 

Table 108 

Statistics of length (1) and length of dental plates (dl) in 45 pedicle valves of Rafinesquina delicata 

sp. nov. 

1 mm (var 1) 11-16(16-566) 

dl mm (var dl) 1-03 (0-124) 

r 0-864 

a (var a) 0-0867 (0-00004) 

Table 109 

Statistics of length (1) and length of ventral muscle scar (sc) in 25 pedicle valves of Rafinesquina 

delicata sp. nov. 

1 mm (var 1) 12-4 (10-79) 

s"c mm (var sc) 3-51(1-804) 

r 0-916 

logel ( var logel) 2-4839 (0-0677) 

logesc (var log e sc) 1-1867 (0-1366) 

r e 0923 

a (var a) 1-4209 (0-01296) 



ordovician brachiopoda 145 

Table iio 

Statistics of length (1) and width (w) of the ventral muscle scar in 23 pedicle valves of Rafinesquina 

delicata sp. nov. 

Imm(varl) 3-6 (1-351) 

w mm (var w) 3-56 (1-107) 

r 0-96 

a (var a) 0-9052 (0-00302) 



KJAERINA (HEDSTROEMINA) Bancroft 1929 
Kjaerina (Hedstroemina) sp. 

(PI. 26, fig. 2) 

The internal mould of a pedicle valve (BB 35452) from the Spy Wood Grit cropping 
out 1 100 yds NNE of Rorrington (Grid Ref. SJ 303015) will probably be shown to 
represent a new species of the Kjaerina species group known as Hedstroemina 
(Bancroft 1929 : 56). The convex valve was about 20 mm long and an estimated 
30 mm wide with roundly obtuse cardinal angles affected by a strong but sharply 
obtuse geniculation to define a relatively flat disc about 10-5 mm long and a trail 
15 mm long. Fragments of the external mould at the antero-medial margin 
indicated that the radial ornamentation was coarsely unequally parvicostellate with 
about 4 ribs per mm. The ventral interarea was strongly apsacline, the foramen 
large and supra-apical with a small or resorbed pseudodeltidium. The well-developed 
teeth were supported by divergent dental plates subtending an angle of about ioo° 
and extending forward for about 15% of the length of the disc. Low convergent 
ridges extended from the anterior ends of the dental plates to enclose a subcircular 
muscle field about 70% as long as the disc with a broad median adductor scar extend- 
ing anteriorly as far as the subflabellate diductor impressions. 

The specimen is quite distinct from other known Hedstroemina in the strong 
geniculation of the valve and the large size of the scar in relation to the disc. The 
Actonian H. robusta (Bancroft 1929 : 59) and H. holtedahli (Spjeldnaes 1957 : 131) 
from the 4ba-/? horizons of Norway compare most closely, but neither is so sharply 
deflected nor characterized by so small a disc. 



Subfamily GLYPTOMENINAE Williams 1965 
BYSTROMENA gen. nov. 

Name. A strophomenacean with the posterior of the adult pedicle valve blocked 
by a plug (Gk. flvoTpa). 

Diagnosis. Concavo-convex, semi-elliptical strophomenids with large protegula, 
ornamented by parvicostellae which may be unequaUy developed in early growth 
stages ; ventral and dorsal interareas apsacline and anacline respectively, pseudo- 
deltidium bounding a large supra-apical foramen truncating the ventral beak, 
arched chilidium relatively large. 



146 SHELVE DISTRICT 

Ventral interior with obliquely placed teeth supported by short widely divergent 
dental plates ; umbonal region of mature valves occupied by a short cylindroid 
plug of secondary shell with a median groove on the dorsal surface ; ventral muscle 
scar obscure, floors of mature valve with many radiating partitions suggesting a 
lemniscate mantle canal system. 

Dorsal interior with delicate cardinalia consisting of a small bilobed cardinal 
process with a vestigial median ridge present between the lobes in young specimens, 
each lobe continuous with widely divergent socket ridges ; notothyrial platform 
absent in young specimens so that the postero-medial part of the valve floor enclosed 
by the cardinalia forms a hollow which is filled in mature valves by a low median 
ridge bifurcating anteriorly ; muscle scars obscure with two pairs of transmuscle 
septa identifiable in young specimens but masked in adult valves by strongly devel- 
oped radiating partitions associated with a lemniscate mantle canal system. 

Type species. Bystromena perplexa sp. nov. from the Spy Wood Grit of the 
Shelve area, Shropshire. 

Discussion. The shape of the shell, the disposition of the teeth and their supports 
and the delicacy of the cardinalia indicate that the new genus is closely related to the 
Glyptomeninae (Williams in Williams et al. 1965 : H388). Yet other features 
immediately distinguish Bystromena from known members of that subfamily. They 
include the tendency for the ribbing to be costellate rather than unequally parvi- 
costellate, although the general pattern is like that of Platymena (Cooper 1956 : 879), 
and the development of radiating partitions in the interior of adult shells. How- 
ever, the most important distinction is undoubtedly the cylindroid mass of secondary 
shell plugging the postero-medial part of the beak interior in adult shells. The 
plug is in striking contrast to the well-developed foramen in young valves. There 
are, moreover, indications that the anterior surface of the plug forms a hollow which 
suggests that the structure was an internal pedicle tube, like those found in some 
triplesiaceans (Wright 1963 : 744), but filled in with secondary shell following atrophy 
of the pedicle. 

Bystromena perplexa sp. nov. 
(PI. 25, fig. 14 ; PI. 26, figs. 3-14 ; PI. 27, figs. 1, 2 ; Text-fig. 11) 

Diagnosis. Transversely semi-elliptical Bystromena with a brachial valve about 
three-quarters as long as wide and gently concave umbonally flattening peripherally 
in maturity, and an evenly convex pedicle valve about one-tenth as deep as long ; 
radial ornamentation finely costellate in adult shells with a modal count of 7 per mm, 
5 mm antero-medially of the umbo. 

Description. Transversely semi-elliptical Bystromena with slightly acute to 
obtuse cardinal angles, brachial valve about three-quarters as long as wide, slightly 
concave in the umbonal region but flattening peripherally in adult growth stages, 
pedicle valve evenly convex and about one-tenth as deep as long, protegulum large, 
trilobed up to 1 mm long ; radial ornamentation commonly unequally parvicostellate 
in early growth stages but becoming costellate in adult shells with counts of 5, 6, 7 



ORDOVICIAN BRACHIOPODA 



147 




cardinal process 

-tooth socket 

\ . ridge 

•O* — P lu 9 

muscle scar 




A B 

Fig. 11. Diagrammatic views of (A) the ventral and (B) the dorsal interiors of 

Bystromena, 



and 8 ribs per mm, 5 mm antero-medially of the umbones, of 3, i, 4 and 1 valves 
respectively ; ventral and dorsal interareas apsacline and anacline respectively with 
a functional foramen at least in young pedicle valves, and well-developed pseudo- 
deltidium and chilidium. 

Teeth obliquely disposed ridges supported by widely divergent dental plates 
extending anteriorly for less than one-tenth of the pedicle valve ; ventral muscle 
field obscure, cylindroid infilled pedicle tube, with a shallow median furrow on the 
dorsal surface, extending anteriorly for about 1 mm ; floor of mature valve orna- 
mented by many radiating thin partitions. 

Cardinal process small with vestigial median ridge present between the lobes in 
immature valves, lobes continuous with widely divergent socket ridges subtending 
an angle of about 135 ° ; notothyrial platform absent except in mature valves when 
it extended forward for a short distance as a median ridge bifurcating anteriorly ; 
transmuscle septa present but usually obscured by radiating thin partitions delineat- 
ing a pinnate mantle canal system ; dorsal muscle impressions obscure. 

Type material. 



Holotype Internal mould of pedicle valve (BB 35363) 
Paratypes External and internal moulds of incomplete 

pedicle valve (BB 35364a, b) 

External and internal moulds of pedicle valve 

(BB 35365a, b) 

External and internal moulds of pedicle valve 

(BB 35366a, b) 

External and internal moulds of brachial valve 

(BB 35367a, b) 

External and internal moulds of brachial valve 

(BB 35368a, b) 

Incomplete internal mould of pedicle valve 

(BB 35369) 

Incomplete internal mould of brachial valve 

(BB 35370) 



length 
8-0 



6-o 



width (mm) 
"•5 



8-o 



4-5 



5'0 



4-0 



6-5 



148 SHELVE DISTRICT 

Type horizon and localities. Spy Wood Grit : BB 35363, 35365 from ex- 
posures on top of ridge 1440 yds NNE of Rorrington bench mark 599 (Grid Ref. 
SJ 303018) ; BB 35364, 35370 from outcrops 1100 yds NNE of Rorrington (Grid 
Ref. SJ 303015) ; BB 35368, 35369 from the north bank of Spy Wood Brook, 170 yds 
north-east of Spy Wood Cottage (Grid Ref. SO 282958) ; BB 35366, 35367 from 
exposures 700 yds NNW of Middleton Church (Grid Ref. SO 296998). 

Discussion. As far as is known, no previously described species can be assigned 
to Bystromena ; and even B. perplexa is represented by only a few moulds, mostly 
imperfectly preserved, from the Spy Wood Grit so that much remains to be discovered 
about the variability of the new species itself. 



Family LEPTAENIDAE Hall & Clarke 1894 

LEPTAENA Dalman 1828 

Leptaena cf. ventricosa Williams 

(PI. 27, figs. 5, 8, 9) 

Diagnosis. Leptaena with a semicircular disc about 10 mm long, ornamented 
by relatively coarse costellae numbering 4 per mm, 5 mm antero-medially of the 
umbo and about 9 low concentric rugae on the disc with a wavelength of 1 mm, sub- 
circular ventral muscle scar less than two-thirds as long as the disc. 

Description. Concavo-convex, sharply geniculate and subquadrate Leptaena 
with a semicircular disc about 10 mm long, almost twice as wide and about one- 
eighth as deep in the vicinity of the ventral umbo, well defined by a confining fold 
with a wavelength of 2 mm, trail sharply deflected at about 90 to give an overall 
depth of about half the length of the pedicle valve ; ventral and dorsal interareas 
apsacline and anacline respectively with a small pseudodeltidium truncated by a large 
open supra-apical foramen and a well-developed chilidium ; radial ornamentation 
finely but evenly costellate with 4 per mm, 5 mm antero-medially of the umbo ; 9 
more or less continuous concentric rugae with a wavelength of about 1 mm also 
ornament the disc. 

Teeth obliquely disposed and supported by short widely divergent dental plates ; 
ventral muscle field not well preserved but apparently rhomboidal in outline and 
about three-fifths as long as the disc. 

Cardinalia poorly preserved but a strong bilobed cardinal process, widely splayed 
socket ridges and an anchor-shaped notothyrial platform were identifiable, as were 
the transmuscle septa especially the longer, curved submedial pair which extend 
anteriorly for about four-fifths of the length. 

Figured material. 

length width (mm) 
External and internal moulds of brachial valve (BB 35453a, b) - - 

Internal mould of pedicle valve (BB 35454) - - 

Horizon and localities. Whittery Shales : BB 35454 from exposures in the 
stream at the north end of Spring Coppice, 865 yds south-east of Hockleton Bridge 



ORDOVICIAN BRACHIOPODA 149 

(Grid Ref. SO 279997) ; BB 35453 from exposures half-way down the path going 
southwards below Marrington Farm (Grid Ref. SO 272967). 

Discussion. The few impressions of Leptaena found in the Whittery and Hagley 
Shales compare very closely with those of L. ventricosa Williams (1963 : 462) from the 
Gelli-grin Group of the Bala district. The disc of L. ventricosa s.s., which is about 
three-fifths as long as wide, as well as the ventral muscle scar, may prove to be 
consistently longer ; but provisionally the Shropshire material is reasonably identi- 
fied as conspecific with the Welsh. 

KIAEROMENA Spjeldnaes 1957 

Kiaeromena cf. kjerulfl (Holtedahl) 

(PI. 27, figs. 6, 10, 11) 

1916 Leptaena kjerulfi Holtedahl : 72. 

1957 Kiaeromena kjerulfi (Holtedahl) Spjeldnaes : 183. 

1963 Kiaeromena cf. kjerulfi (Holtedahl) Williams : 464. 

Diagnosis. Obtusely geniculate Kiaeromena with a disc about 12 mm long, 
ornamented by up to 9 coarse concentric rugae and unequally developed parvicostel- 
lae numbering about 6 per mm, 10 mm antero-medially of the ventral umbo ; 
subcircular ventral muscle field about 6 mm long. 

Description. Subquadrate Kiaeromena with a rounded obtuse angle of genicula- 
tion to define a plano-convex disc about 12 mm long and almost twice as wide 
ornamented by up to 9 coarse concentric asymmetric rugae with wavelengths of 
1-5 mm and steeper posterior slopes ; radial ornamentation unequally parvicostellate 
especially on the trail with a count of 6 per mm, 10 mm antero-medially of the ventral 
umbo ; ventral and shorter dorsal interareas apsacline and anacline respectively 
with the pseudodeltidium and beak pierced by a large supra-apical foramen, and an 
arched chilidium. 

Ventral interior with trigonal teeth supported by divergent dental plates acting 
as the posterior boundaries of a poorly preserved subcircular muscle scar extending 
forward for about 6 mm. 

Dorsal interior with a cardinal process consisting of a small median ridge flanked 
by two acutely divergent lobes and poorly defined socket ridges supported by a 
massive, anchor-shaped notothyrial platform. 

Figured material. 

length width (mm) 
Incomplete external and internal moulds of pedicle valve 
(BB 35406a, b) 25 

Incomplete external and internal moulds of brachial valve 
(BB 37132a, b) - - 

Horizons and localities. BB 35406 from Whittery Shales exposed in the lower 
part of the west bank of the River Camlad, 60 yds north-east of Marrington Farm 
(Grid Ref. SO 272970) ; BB 37132 from Hagley Volcanics exposed in a roadside 
quarry 150 yds north of Church Stoke Hall (Grid Ref. SO 274942). 



150 SHELVE DISTRICT 

Discussion. Only a few incomplete moulds of a finely ribbed, obtusely geniculate 
leptaenid have been recovered from the Hagley and Whittery Shales. There is 
no doubt, however, about the generic identity of the specimens which compare 
sufficiently closely with Kiaeromena cf. kjerulfi, occurring rarely in the Allt Ddu 
Group of the Bala district (Williams 1963 : 464), to be regarded as conspecific. 

Order PENTAMERIDA Schuchert & Cooper 1931 

Suborder SYNTROPHIIDINA Ulrich & Cooper 1936 

Superfamily PORAMBONITACEA Davidson 1853 

Family SYNTROPHIIDAE Schuchert 1896 

Subfamily XENELASMATINAE Ulrich & Cooper 1936 

EUORTHISINA Havlicek 1950 

The genus Euorthisina Havlicek (1950 : 16) was not included by Biernat in her 
review of the Porambonitacea (in Williams et at. 1965 : H523-536). But, as demon- 
strated by Havlicek, the genus is undoubtedly a porambonitacean although unique 
in its shape within the superfamily. This difference is related to the strong devel- 
opment of wide interareas in both valves and the absence of a dorsal fold and ventral 
sulcus. Such features are normally quite variable in other articulates but they are 
so exceptional among the porambonitaceans as to prompt Biernat (1965 : H524) to 
report that within the superfamily the interareas are 'commonly reduced and scarcely 
visible' and that the 'anterior commissure is invariably uniplicate'. Neither 
diagnosis is true of Euorthisina which is, accordingly, immediately distinguishable 
from all other porambonitaceans. Through the kindness of Dr Havlicek, it is 
possible to figure here internal moulds of Euorthisina moesta (Barrande) from the 
Llanvirnian Sarka formation of Rokycany (PI. 28, figs. 4, 6, 7) which show the main 
features of the genus (BB 37160-37162). 

Havlicek (1950 : 16) compared his genus with Xenelasma (Ulrich & Cooper 
1936 : 631) which is also characterized by discrete dental plates and a septalium. It 
is, however, possible that the internal similarity is an expression of convergence, 
but until more is known about the ancestry of Euorthisina it seems more practicable 
to follow Havlicek's lead and provisionally retain the genus within the Xenelasmat- 
inae. 

Euorthisina cf. moesta minor Havlicek 
(PL 28, figs. i-3, 5, 8) 

1950 Euorthisina moesta minor Havlidek : 87. 

Diagnosis. Subequally and evenly biconvex, transversely suboval Euorthisina 
ornamented by fine lamellae and costellae numbering 3 per mm, 5 mm antero- 
median^ of the ventral umbo ; parallel dental plates and septalium extending for- 
ward for about one-fifth the length of their respective valves. 

Description. Transversely suboval Euorthisina with a gently convex brachial 
valve about 60% as long as wide and a pedicle valve about 15% as deep as long with 



ORDOVICIAN BRACHIOPODA 151 

an evenly and gently convex transverse profile ; ornamented by fine closely spaced 
lamellae and rounded ribs composed of 9 primary costae and intercalated costae 
within 1 mm of the umbo and costellae appearing in later stages of growth by branch- 
ing and intercalation with counts of 3 ribs per mm, 5 mm antero-medially of the 
umbones of 2 pedicle valves ; ventral interarea short, almost orthocline, with a narrow 
open delthyrium ; brachial interarea not exposed. 

Ventral interior with simple node-like teeth supported by long parallel dental 
plates extending forward for an average of one-fifth the length of 3 pedicle valves to 
define a narrow delthyrial chamber ; muscle and mantle canal impressions unknown. 

Dorsal interarea with small septalium supported by a median ridge extending 
forward for about one-fifth the length of the brachial valve ; muscle and mantle 
canal impressions unknown. 

Figured material. , ,, . ,,, , . 

length width (mm) 

External and internal moulds of orthocone containing im- 
pressions of 4 Euorthisina (BB 35377a, b) 
Incomplete external and internal moulds of pedicle valve 
(BB 35378a, b) 4-5 

External and internal moulds of pedicle valve (BB 35379a, b) 3-2 - 

Incomplete external and internal moulds of brachial valve 
(BB 35380a, b) 4-5 7-5 

Horizon and localities. Mytton Flags : BB 35377, 35378 from the adit in 
Maddox's Coppice, 1200 yds ENE of St Luke's Church, Snailbeach (Grid Ref. 
SJ 382030) ; BB 35379, 35380 from outcrops by the side of the cart-track 140 yds 
north of Wood House, Gravels (Grid Ref. SJ 338003). 

Discussion. The small number of moulds of Euorthisina collected from the 
Mytton Flags include a group of 4 poorly preserved conjoined shells all orientated 
in the same way within an orthocone and probably buried in position of growth. 
The sample compares most closely with E. moesta minor Havlicek from the Llan- 
virnian Sarka Shales of Czechoslovakia, especially in the relative dimensions and 
size of the dental plates (cf. PI. 28, figs. 4, 6, 7). Havlicek (1950 : 87) reports that 
in his subspecies there are 8 to 12 ribs in the umbonal region compared with 14 to 16 
in the Shelve specimens. But this difference may be an expression of variability 
in the older stock and its systematic importance can only be satisfactorily assessed 
when more material is available. 

Family PARASTROPHINIDAE Ulrich & Cooper 1938 

PARASTROPHINELLA Schuchert & Cooper 1931 

Parastrophinella musculosa sp. nov. 

(PL 28, figs. 9-13, 17) 

Diagnosis. Small, subtriangular, strongly biconvex Parastrophinella with vesti- 
gial fold and sulcus and commonly 11 to 13 costae, with a wavelength of 0-6 mm, 
ornamenting the anterior part of the shell. 



152 SHELVE DISTRICT 

Description. Small, biconvex, subtriangular Parastrophinella with a brachial 
valve averaging 93% as wide as long and 29% as deep as long (for 4 and 3 valves 
respectively) and a mean depth of 23% relative to the length of 3 pedicle valves, 
transverse profile evenly convex with steep lateral slopes, longitudinal profile evenly 
convex to anteriorly geniculate, dorsal fold and ventral sulcus faintly and sporadically 
developed ; rostrate with depressed palintropes extending antero-laterally for up to 
35% OI * ne maximum width of the shell, delthyrium and notothyrium open ; shell 
smooth for an average of 37 mm anterior of the umbones of 3 specimens, delayed 
costation consisting of rounded costae with a mean wavelength of 0-56 mm between 
5 and 6 mm anterior of the umbones of 7 brachial valves and numbering 10 to 14 in 
1, 2, 1, 2 and 1 brachial valves respectively between 5 and 7 mm long. 

Ventral interior with small teeth and well-developed spondylium sessile posteriorly 
but supported anteriorly by a median septum extending forward for an average of 
45% °f the length of 3 pedicle valves. 

Dorsal interior with subparallel outer plates separated by less than one-tenth the 
maximum width of the brachial valve and extending forward for a mean proportion 
of 40% of the length of 4 valves ; elongately oval adductor scars impressed on either 
side of the outer plates for an average of 56% of the length of 3 valves. 

Type material. 

length width (mm) 
Holotype External and internal moulds of pedicle valve 

(BB 35598a, b) 7-0 8-5 
Paratypes External and internal moulds of brachial valve 

(BB 35599a, b) 6-o 5-8 

Internal mould of brachial valve (BB 35600) 5-0 5-2 

Internal mould of pedicle valve (BB 37107) 5-5 5-0 

Internal mould of brachial valve (BB 37108) 7-5 7-0 

Type horizon and locality. Spy Wood Grit exposed 1100 yds NNE of 
Rorrington (Grid Ref. SJ 303015). 

Discussion. The Spy Wood Grit porambonitacean is an early species of Para- 
strophinella which is like the Upper Llandeilo P. costata MacGregor (1961 : 199) 
except for its significantly fewer, coarser ribs, its relatively greater length and the 
vestigial nature of its fold and sulcus. These differences are important enough to 
merit systematic recognition. 

Parastrophinella sp. 

(PI. 28, figs. 14, 15) 

Incomplete external and internal moulds (BB 35461a, b) of a pedicle valve, 
collected from the Hagley Volcanics in the roadside quarry 150 yds. north of Church 
Stoke Hall (Grid Ref. SO 274942), probably represent Parastrophinella. The valve, 
which was 13-5 mm long, was 85% as long as wide and 25% as deep as wide with a 
persistent flat-bottomed sulcus about 60% as wide as long and evenly convex flanks. 
Radial ornamentation consisted of rounded costae about 0-8 mm in wavelength, 



ORDOVICIAN BRACHIOPODA 153 

5 mm antero-medially of the umbo where they arose, and numbered 4 in the sulcus 
and 8 on each flank. The apsacline interarea was relatively short and curved, and 
the open delthyrium was flanked by small teeth supported by the walls of a sessile 
spondylium extending anteriorly for 37% of the length of the valve. 

In the absence of a complementary brachial valve, the identification of the remains 
of the pedicle valve as Parastrophinella is provisional. Furthermore, the valve is 
quite different from older Parastrophinella found in Shropshire and Wales with their 
weak sulci and smooth umbonal regions. More material is, therefore, required not 
only to indicate the specific affinities of the stock but also to verify the generic status 
of the specimen. 



Order RHYNCHONELLIDA Kuhn 1949 

Superfamily RHYNCHONELLACEA Gray 1849 

Family TRIGONIRHYNCHIIDAE McLaren 1965 

ROSTRICELLULA Ulrich & Cooper 1942 

Rostricellula sparsa Williams 

(PI. 27, figs. 3, 4, 7) 

1963 Rostricellula sparsa Williams : 467. 

Diagnosis. Small, relatively wide Rostricellula with 2 costae, averaging 0-4 mm 
thick 2 mm anterior of the umbo, on a well-developed fold two-fifths as wide as the 
length of the brachial valve, and 4 to 6 costae on each of the lateral slopes. 

Description. Small, biconvex, subtriangular Rostricellula with a brachial valve 
slightly less long than wide (96% for 2 valves) and 15% as deep as long, and a pedicle 
valve slightly less than one-quarter as deep as long, dorsal fold (and ventral sulcus) 
strongly developed up to 40% as wide as the length of the brachial valve and rounded 
in transverse profile with evenly convex lateral slopes ; rostrate with high curved 
ventral umbo and inconscpicuous dorsal umbo, apical angle about 100°, nature of 
deltidial plates unknown ; radial ornamentation exclusively costate with 2 rounded 
costae on the fold, averaging 0-4 mm in wavelength, 2 mm anterior of the umbones of 
4 brachial valves, and 4 to 6 on the lateral slopes. 

Ventral interior with small teeth supported by slightly divergent dental plates 
extending anteriorly for about one-eighth the length of the valve ; muscle field 
unknown. 

Hinge-plate of dorsal interior small, divided by septalium extending forward for 
an average of 8% the length of 4 brachial valves, median ridge low, rounded but 
developed for about two-fifths the length of the valve ; details of crura and muscle 
scar unknown. 

Figured material. 

length width (mm) 

External and internal moulds of brachial valve (BB 35584a, b) 3-5 3-5 

External and internal moulds of pedicle valve (BB 35585a, b) 3-0 - 



154 SHELVE DISTRICT 

Horizon and locality. Spy Wood Grit exposed noo yds NNE of Rorrington 
(GridRef. S J 303015). 

Discussion. Rostricellula is rare in the Shelve area, being represented only by a 
few valves from the Spy Wood Grit which are quite different in ornamentation and 
outline from all other described species except R. sparsa from the Allt Ddu Mud- 
stones of the Bala district. In every morphological aspect the Spy Wood specimens 
are identical with the small sample on which the Welsh species was based. Especially 
significant for immediate identification is the presence of 2 costae on the fold. Only 
one American and one European species are known with so few ribs. They are : 
R. ? minuta from the Hermitage Formation of Tennessee which, although similar in 
shape, has 7 or 8 costae on each lateral slope (Cooper 1956 : 640) ; and R. ambigena 
(Barrande) from the Upper Ordovician of Bohemia (Havlicek 1961 : 51) which is not 
only relatively longer but has the costae on the fold and sulcus branching within the 
size range of the British species. 



Order SPIRIFERIDA Waagen 1883 

Suborder ATRYPIDINA Moore 1952 

Superfamily ATRYPACEA Gill 1871 

Family ATRYPIDAE Gill 1871 

Subfamily ZYGOSPIRINAE Waagen 1883 

ZYGOSPIRA Hall 1862 

Zygospira sp. 

(PI. 27, figs. 12, 13) 

An internal mould (BB 37133) has been recovered from Spy Wood Grit outcrops 
1 100 yds NNE of Rorrington (Grid Ref. SJ 303015), which is likely to represent the 
remains of a brachial valve of the atrypacean Zygospira. The rostrate valve, which 
was about 2 mm long, was nearly circular in outline and about 15% as deep with a 
shallow narrow median sulcus and evenly convex lateral areas. The external surface 
was ornamented by about 15 fine costae while internally a median ridge corresponding 
to the narrow sulcus supported a divided hinge plate with rounded crural bases. 

This occurrence is one of the earliest records of Zygospira, which has hitherto 
been described from younger Ordovician strata in N. America (Cooper 1956 : 672- 
674), Scotland (Williams 1962 : 242) and Mid-Europe (Havlicek & Vanek 1966 : 61). 
The relatively few ribs may be diagnostic but may also reflect immaturity of the 
specimen and detailed comparison with established species would not be profitable 
at present. 

VI. ACKNOWLEDGMENTS 

This systematic survey could not possibly have been undertaken without the active 
encouragement of the late Professor W. F. Whittard, F.R.S. who so generously put 
his unique collection of Shelve brachiopods at my disposal and indeed sponsored 



ORDOVICIAN BRACHIOPODA 155 

further collecting in certain localities by Mr T. R. Fry in the early stages of my 
researches. The study is, accordingly, dedicated to his memory ; and whatever 
worth its comprehensiveness may have is a tribute to his inspiring devotion to 
Shropshire Geology. 

I wish to record my thanks to Dr W. T. Dean of the Geological Survey of Canada 
who selflessly accepted responsibility for preparing a geological map of the Shelve 
area from Professor Whittard's field sheets, and for spending so much time in 
providing me with precise locations for nearly all the specimens in the collection. 
The text has greatly benefited from discussions with Dr A. D. Wright and the 
illustrations and statistical tables from the assistance of Dr Jean Graham, both of 
the Queen's University, Belfast. I am also indebted to Dr V. Havlicek for sending 
me specimens of Euorthisina and to Dr L. R. M. Cocks and Dr G. F. Elliott of the 
British Museum (Natural History) for their help in a number of matters relating to 
literature and the layout of the paper. 



VII. REFERENCES 

Alichova, T. N. 1953. A guide to the brachiopod fauna in the Ordovician sediments of the 

north-western part of the Russian Platform. Trudy vses. nauchno-issled. geol. Inst., 

Leningrad, 1953 : 1-127, pis. 1-17. [In Russian.] 
Andreeva, O. N. i960. Novye vidy drevnikh rasteniy i Bezprovochnykh S.S.S.R. [Markowskii, 

B. P., editor-in-chief], 1 : 288 pp., 73 pis. Moscow. 
Bancroft, B. B. 1928. The Harknessellinae. Mem. Proc. Manchr lit. phil. Soc, 72 : 173-196, 

pis. 1-2. 

1929. Some new genera and species of Strophomenacea from the Upper Ordovician of 

Shropshire. Mem. Proc. Manchr lit. phil. Soc, 73 : 33-65, pis. 1-2. 

1945- The brachiopod zonal indices of the stages Costonian to Onnian in Britain. /. 

Paleont., Menasha, 19 : 181-252, pis. 22-38. 

Barrande, J. 1848. Uber die Brachiopoden der Silurischen Schichten von Boehmen, 2 : 153-256, 
pis. 15-23. Vienna. 

1879. Systeme Silurien du centre de la Boheme, 5. Classe des Mollusques. Ordre des 

Brachiopodes . 226 pp., 153 pis. Prague. 

Bassett, D. A., Whittington, H. B. & Williams, A. 1966. The stratigraphy of the Bala 

district, Merionethshire. Q. J I geol. Soc. Lond., 122 : 219-271. 
Bates, D. E. B. 1968. The Lower Palaeozoic brachiopod and trilobite faunas of Anglesey. 

Bull. Br. Mus. nat. Hist. (Geol.), London, 16 : 127-199, pis. 1-14. 
1969. Some early Arenig brachiopods and trilobites from Wales. Bull. Br. Mus. nat. 

Hist. (Geol.), London, 18 : 1-28, pis. 1-9. 
Cockerell, T. D. A. 1911. The name Glossina. Nautilus, Philad., 25 : 96. 
Cooper, G. A. 1956. Chazyan and related brachiopods. Smithson. misc. Collns, Washington, 

127 : 1-1024, 1025-1245, pis. 1-269. 
Davidson, T. 1868. British Fossil Brachiopoda, 3. Devonian and Silurian Brachiopoda. 

Palaeontogr. Soc. (Monogr.), London, 397 pp., 50 pis. 

1883. Supplement to the Fossil Brachiopoda. Palaeontogr. Soc. (Monogr.), London, 

242 pp., 17 pis. 

Dean, W. T. 1958. The faunal succession in the Caradoc Series of south Shropshire. Bull. 
Br. Mus. nat. Hist. (Geol.), London, 3 : 191 -231, pis. 24-26. 

Goryansky, V. 1969. Inarticulate Brachiopods of the Cambro-Ordovician rock of the north- 
west Russian Platform. 126 pp., 21 pis. NEDRA, Leningrad. [In Russian.] 



156 SHELVE DISTRICT 

HAVLfCEK, V. 1949. Orthoidea and Clitambonoidea of the Bohemian Tremadoc. Sb. st. 
geol. Ust. csl. Repub., Praha, 16 : 93-144, pis. 1-5. 

1950. The Ordovician Brachiopoda from Bohemia. Rozpr. ustred. Ust. geol., Praha, 

13: 1-135. Pis- 1-13- 

1961. Rhynchonelloidea des Bohmischen Alteren Palaozoikums (Brachiopoda). Rozpr. 

ustred. Ust. geol., Praha, 27 : 1-211, pis. 1-27. 

1967. Brachiopoda of the suborder Strophomenidina in Czechoslovakia. Rozpr. ustred. 

Ust. geol., Praha, 33 : 1-235, pis. 1-52. 

1968. New brachiopods in the Lower Caradoc of Bohemia. Vest, ustred. Ust. geol., 

Praha, 43 : 123-125, pis. 1-2. 

1970. Heterorthidae (Brachiopoda) in the Mediterranean Province. Sb. geol. Ved. 

Praha, (P) 12 : 7-39, pis. I-II. 

- 1971. Brachiopodes de l'Ordovicien du Maroc. Notes M6m. Serv. Mines Carte gdol. 

Maroc, Rabat, 230 : 1-135, pis. 1-26. 
& Vanek, J. 1966. The biostratigraphy of the Ordovician of Bohemia. 56. geol. Ved. 

Praha, (P) 8 : 7-69, pis. 1-16. 
Holtedahl, O. 1916. The Strophomenidae of the Kristiania Region. Skr. VidenskSelsk. 

Christiania, 12 : 1-117, pis. 1-16. 
Jones, O. T. 1928. Plectambonites and some allied genera. Mem. geol. Surv. Gr. Br.Palaeont., 

London, 1 (5) : 367-527, pis. 21-25. 
Kozlowski, R. 1929. Les brachiopodes gothlandiens de la Podolie Polonaise. Palaeont. 

pol., Warsaw, 1 : 1-254, pl s - 1-12. 
Lapworth, C. & Watts, W. W. 1894. The geology of South Shropshire. Proc. Geol. Ass., 

London, 13 : 297-355, pis. 8-9. 
MacGregor, A. R. 1961. Upper Llandeilo brachiopods from the Berwyn Hills, North 

Wales. Palaeontology, London, 4 : 177-209, pis. 19-23. 
Mickwitz, A. 1909. Vorlaufige Mitteilung iiber das Genus Pseudolingula Mickwitz. M6m. 

Acad. Sci. St. Petersb. (6) 3 : 765-772. 
Murchison, R. I. 1839. The Silurian System [&c.]. 768 pp., 37 pis. London. 

1859. Siluria. 3rd ed. 592 pp., 42 pis. London. 

Neuman, R. B. 1964. Fossils in Ordovician tuffs, Northeastern Maine. Bull. U.S. geol. 

Surv., Washington, 1181-E : 1-38, pis. 1-7. 

1968. Paleogeographic implications of Ordovician shelly fossils in the Magog belt of the 

Northern Appalachian Region. In Zen, E-an, et al. (eds). Studies of Appalachian 
Geology : Northern and Maritime : 35-48, pis. New York. 

Nikiforova, O. I. & Andreeva, O. N. 1961. Ordovician and Silurian stratigraphy of the 

Siberian Platform and its palaeontological basis (Brachiopoda). Trudy vses. nauchno- 

issled. geol. Inst., Leningrad, 56 : 1-412, pis. 1-56. [In Russian.] 
Opik, A. 1930. Brachiopoda Protremata der Estlandischen Ordovizischen Kukruse-Stufe. 

Tartu Ulik. Geol.-Inst. Toim., 20 : 1-261, pis. 1-22. 

1934- Uber Klitamboniten. Tartu Ulik. Geol.-Inst. Toim., 39 : 1-239, pis. 1-48. 

1939. Brachiopoden und Ostrakoden aus dem Expansusschiefer Norwegens. Norsk 

geol. Tidsskr., Oslo, 19 : 117- 142, pis. 1-6. 
Phillips, J. 1848. The Malvern Hills, compared with the Palaeozoic districts of Abberley, 

Woolhope, May Hill, Tortworth and Usk. Mem. geol. Surv. U.K., London, 2 : 1-330, 

pis. 1-3. 
Roomusoks, A. K. 1959. Strophomenoidea Ordovika i Silura Estonii 1. Rod Sowerbyella 

Jones. Tartu Univ. (Dorpat), Act. Commentationes 75 : 11-41, pis. 1-8. 
1970. Stratigraphy of the Viruan Series (Middle Ordovician) in Northern Estonia. Tallin, 

U.S.S.R. 345 pp., 8 pis. [In Russian.] 
Rowell, A. J. 1966. Revision of some Cambrian and Ordovician inarticulate brachiopods. 

Paleont. Contr. Univ. Kans., Topeka, 7 : 1-36, pis. 1-4. 



ORDOVICIAN BRACHIOPODA 157 

Rubel, M. P. 1961. Lower Ordovician brachiopods of the superfamilies Orthacea, Dalmanel- 
lacea and Syntrophiacea of the eastern Baltic. Eesti NSV Tead. Akad. geol. Inst. Uuri- 
mused, 6 : 141-226, pis. 1-27. [In Russian.] 

Rudwick, M. J. S. 1965. Sensory spines in the Jurassic brachiopod Acanthothiris. Palaeont- 
ology, London, 8 : 604-617, pis. 84-87. 

Salter, J. W. 1859. Appendix A. In Murchison, R. I. Siluria, 3rd ed. : 531-552. London. 

1866. Appendix. On the fossils of North Wales. In Ramsey, A. C. The Geology of 

North Wales. Mem. geol. Surv. U.K., London, 3 : 239-381, pis. 1-26. 

Schuchert, C. & Cooper, G. A. 1932. Brachiopod genera of the suborders Orthoidea and 
Pentameroidea. Mem. Peabody Mus. Yale, New Haven, 4 : 1-270, pis. 1-29. 

Sinclair, G. W. 1945. Some Ordovician Lingulid Brachiopods. Trans. R. Soc. Can., 
Ottawa, (3, iv) 39 : 55-82, pis. 1-4. 

Sowerby, J. de C. 1839. In Murchison, R. I. The Silurian System [&c] : 608-644, pis. 5-22. 
London. 

Spjeldnaes, N. 1957. The Middle Ordovician of the Oslo Region, Norway. 8. Brachiopods 
of the Suborder Strophomenida. Norsk geol. Tidsskr., Bergen, 37 : 1-214, pis. 1-14. 

Temple, J. T. 1968. The Lower Llandovery (Silurian) brachiopods from Keisley, Westmor- 
land. Palaeontogr. Soc. (Monogr.), London, 58 pp., 10 pis. 

1970. The Lower Llandovery brachiopods and trilobites from Ffridd Mathrafal, near 

Meifod, Montgomeryshire. Palaeontogr. Soc. (Monogr.), London, 76 pp., 19 pis. 

Ulrich E. O. & Cooper, G. A. 1936. New genera and species of Ozarkian and Canadian 
brachiopods. /. Paleont., Tulsa, 10 : 616-631. 

1938. Ozarkian and Canadian Brachiopoda. Spec. Pap. geol. Soc. Am., Baltimore, 

13 : 1-323, pis. 1-58. 

Watts, W. W. 1925. The geology of South Shropshire. Proc. Geol. Ass., London, 36 : 321- 

363- 
Whittard, W. F. 1952. A geology of South Shropshire. Proc. Geol. Ass., London, 63 : 143- 

197. 
1966. The trilobites of the Henllan Ash (Lower Ordovician), Merionethshire. Bull. Br. 

Mus. nat. Hist. (Geol.), London, 11 : 489-505, pis. 1-5. 
Whittington, H. B. & Williams, A. 1955. The fauna of the Derfel Limestone of the Arenig 

District, North Wales. Phil. Trans. R. Soc, London, (B) 238 : 397-430, pis. 38-40. 
Williams, A. 1949. New Lower Ordovician brachiopods from the Llandeilo-Llangadock 

District. Geol. Mag., London, 86 : 161 -174, 226-238, pis. 8, n. 

1953- The geology of the Llandeilo district, Carmarthenshire. Q. Jl geol. Soc. Lond., 

108 : 177-208. 

1962. The stratigraphy and brachiopod faunas of the Barr and Lower Ardmillan Series 

(Caradoc) of the Girvan district, south-west Ayrshire, with descriptions of the Brachiopoda. 
Mem. geol. Soc. Lond., 3 : 1-267, pl s - I_2 5- 

1963. The Caradocian brachiopod faunas of the Bala District, Merionethshire. Bull. 

Br. Mus. nat. Hist. (Geol.), London, 8 : 327-471, pis. 1-16. 

1973- Distribution of brachiopod assemblages in relation to Ordovician palaeogeography. 

In Hughes, N. F. (ed.). Organisms and Continents through Time. Spec. Pap. Palaeont., 

London, 12 : 241-269. 
et al. 1965. Brachiopoda. In Moore, R. C. (ed.). Treatise on Invertebrate Paleontology, 

H. 927 pp., illustr. Kansas. 
et al. 1972. A correlation of Ordovician rocks in the British Isles. Spec. Rep. geol. Soc. 

Lond., 3 : 1-74, text-figs. 
Wilson, A. E. 1913. A new brachiopod from the base of the Utica. Bull. Victoria meml 

Mus., Ottawa, 1 : 81-84, P 1 - 8 - 
Wright, A. D. 1963. The fauna of the Portrane Limestone. I. The inarticulate brachiopods. 

Bull. Br. Mus. nat. Hist. (Geol.), London, 8 : 221-254, pis. 1-4. 



158 



SHELVE DISTRICT 



1964. The fauna of the Portrane Limestone, II. Bull. Br. Mus. nat. Hist. (Geol.), 
London, 9 : 157-256, pis. 1-11. 
- 1 97 1. Taxonomic significance of the pseudodeltidium in triplesiacean brachiopods. 
Palaeontology, London, 14 : 342-356, pis. 62-63. 



INDEX 

New taxonomic names and the page numbers of the principal references are printed in bold 
type. An asterisk (*) denotes a figure. 



Acanthothiris 55 

Acrotretacea 42-4 

Acrotretida 42-9 

Acrotretidae 42-4 

Acrotretidina 42-8 

Actonian 14 

Addison, R. 14, 124 

Aegiromeninae 139-40 

Africa, North n, 13 ; see Morocco 

Aldress Shales 9-11, 12*, 14, 15*, 21, 22 

Alimbella 68-9 

Alimbellidae 68, 69-71 

Allt Dhu Mudstones 14 

Anglesey 12, 14 

Anglo-Welsh successions 8, 11, 13, 16 

Anomalorthinae 115 

Apsotreta 42 

sp. 9, 17, 42-3 ; pi. 6, figs. 14-5 
Arenig 11-2, 14, 15*, 18 
Articulata 49-154 
Astraborthis 7, 68-9, 69* 



17, 69, 70-1 ; pi. 11, figs. 5, 



uniplicata 9, 

6, 9, 11, 14 
Atrypacea 154 
Atrypidae 154 
Atrypidina 154 



Bala 14 

Balclatchie Mudstones 11 

Baltic 11 

Berwyn Dome 13 

Betton Beds 9-13, 12*, 15*, 19 

Bicuspina 13-5, 21-3, 117, 124, 126 

cava 119 

modesta 9, 20-1, 118-21 ; pi. 20, 
figs. 3-8 

multicostellata 119 

spiriferoid.es 119 

subquadrata 9, 22-3, 117-8, 119-21 ; 
pi. 19, figs. 17-9 ; pi. 20, figs. 1-2 
Bohemia 8, 11, 13-4, 16 ; see Czechoslovakia 
brachiopoda 1-158 



Bystromena 7, 21, 145-6, 147*, 148 

perplexa 9, 21, 146-8 ; pi. 25, fig. 14 ; 
pi. 26, figs. 3-14 ; pi. 27, figs. 1-2 

Caeroplecia 7, 14, 22, 121, 122* 

plicata 9, 22-3, 121, 122-4 ; pi. 20, 

figs. 14-6 ; pi. 21, figs. 1-6, 8 
Caradoc 13-6, 15* 
Carlisle, Mrs. H. 65 
Carmarthenshire 12, 14 
Cautleyan 14 
' Cliftonia' dorsata 126 
Clitambonitidina 115 -6 
Cocks, Dr L. R. M. 155 
commonness, relative, of species 16 
Conotreta 43 

stapeleyensis 9, 18, 43-4 ; pi. 6, figs. 
16-21 
continental plates, ancient 8 
Coston Beds, Costonian 14, 15*, 16 
Craniacea 48-9 
Craniidae 48-9 
Craniidina 48-9 

currents, transporting capacity 16 
Czechoslovakia 11, 13, 16 ; see Bohemia 

Dalmanella 13-5, 18-9, 21-2, 89 

elementaria 9, 17, 97-8 ; pi. 16, figs. 2-8 
indica 93 

parva 89-90, 92-7 ; pi. 15, figs. 1-4, 7 
prototypa 108 

salopiensis 9, 19-20, 90-1, 92-7 ; pi. 15, 
figs. 5-6, 8-12 
gregaria 9, 21, 91-2, 93-7 ; pi. 15, 

figs. 13-6, 20 
transversa 9, 22, 92-3, 93-7 ; pi. 15, 
figs. 17-9, 21-4 ; pi. 16, fig. 1 
Dalmanellidae 89-102 
Dean, Dr W. T. 8, 155 
Derfel Limestone 14 
Desmorthis 11, 18, 73 

? sp. nov. 9, 17-8, 73-4 ; pi. 12, figs. 2-4, 8 
Didymograptus bifidus zone, shales 13, no 



INDEX 



159 



murchisoni zone 12 
Dinorthis 23 
Diparelasma 11, 71 

sp. 9, 17, 71 ; pi. 11, figs. 3, 7-8 
Discinacea 44-8 
Discinidae 11, 47-8 
Dolerorthidae 63-7 
Dolerorthinae 63-4 
Dolerorthis 63 

cf. tenuicostata 9, 14, 23, 63-4 ; pi. 10, 
figs- 3. 5. 7. 10, 13 
Drabovia 15, 87 

cf. fascicostata 9, 16, 21, 87-8 ; pi. 14, 
figs. 14-6, 18-9 

fascicostatata 87 
Draboviinae 87-9 

Elkaniidae 38-40 

Elliott, Dr G. F. 155 

Enteletacea 87-115 

environments 16, 18 

Eocramatia 7, 127-8, 128-9, 129* 

dissimulata 9, 18, 129-30 ; pi. 21, figs. 
9, 12, 14-5 ; pi. 22, figs. 1-3, 5-6 

Eocramatiidae 7, 127-8, 128-30 

Euorthisina 11, 18, 150, 155 
moesta 150 ; pi. 28, figs. 4, 6-7 
cf. moesta minor 9, 11, 17, 150-1 ; pi. 28, 

figs- i-3. 5. 8 
expansus shale 115 

faunal associations 16-23 

distribution 9-16 
Ffairfach Grit, Group 13 
Finkelnburgia 85 
Finkelnburgiidae 68, 71-85 
Fry, T. R. 8, 155 
Furcitella 140- 1 

plicata 141 

sp. 9, 23, 140-1 ; pi. 25, figs. 4-5 
Furcitellinae 140-1 

Garn-Wen 13 
Gelidorthis 74 

cf. partita 9, 13, 20, 74-5 ; pi. 12, figs. 

5-7 
Gelli-Grin Calcareous Ashes 14, 58 
Giraldiella partita 74 
Glossellinae 35-8 
Glossina 31 
Glossorthis 68 
Glyptoglossella 30 
Glyptomeninae 145-8 
Glyptorthinae 64-7 



Glyptorthis 14, 21, 64 

assimilis 65 

balclatchiensis 65 

crispa 65 

nantensis 65 

viriosa 10, 20-1, 23, 64-7 ; pi. 10, figs. 6, 
8-9, 11-2, 14-5 ; pi. 11, figs. 1-2, 4 

sp. 20 
Gonambonitacea 115-6 
Gonambonitidae 115 
Graham, Dr Jean 155 
graptolites 8, 13 ; see Didymograptus 

Hagley Shales 9-11, 12*, 14, 15*, 21, 22 

Volcanic Group 15* 
Harknessella 102 

cf. subplicata 10, 14, 21, 102 ; pi. 16, 
fig. 16 
Harknessellidae 102-6 
Harnagian 14, 15* 
Havlicek, Dr V. 155 
Hedstroemina 145 

holtedahli 145 

robusta 145 
Henllan Ashes 12 
Hesperonomia 11, 49 

australis 50 

sp. 10, 12, 17, 49-50 ; pi. 7, figs. 9, 12 
Hesperonomiella 12, 50 

carmalensis 50 
Hesperonomiidae 49-50 
Heterorthidae 106-14 
Heterorthis 13-4, 106-7 

retrorsistria 107 

sp. 10, 20-2, 106-7 ; pi. 17, figs. 10-4 
Hope Shales 9-11, 12*, 15*, 18 
Horderleyella 13-4, 102 

convex a 104 

plicata 103 

cf. plicata 10, 14, 21-2, 102-3 ; pi. 16, 
figs. 17, 19-20 ; pi. 17, fig. 1 

subcarinata 103 

sp. 10, 20, 104 ; pi. 17, figs. 2-3 

Inarticulata 13, 25-49 
invertebrates 8 

Kiaeromena 149 

cf. kjerulfi 10, 14, 149-50 ; pi. 27, figs. 6, 
10-1 

sp. 22-3 
Kjaerina {Hedstroemina) 145 

sp. 10, 21, 145 ; pi. 26, fig. 2 



i6o 



INDEX 



Ktaoua Formation, Morocco 14 
Kullervo 115 

panderi 116 

sp. 10, 20, 115-6 ; pi. 19, figs. 14-5 
Kullervoidae 115-6 

lahars 21 
Lenorthis 18, 52 

parvicrassicostatus 53 

proava 12, 53-4 

cf. proava 10-1, 17, 52-3, 55-7 ; pi. 8, 
figs. 2-9 
Leptaena 23, 148 

kjerulfi 149 

sericea 134 

cf. ventricosa 10, 14, 22-3, 148-9 ; pi. 27, 
figs. 5, 8-9 
Leptaenidae 148-50 
Leptestiidae 126-7 
Leptestiinae 126-7 
Letna Formation, Bohemia 15-6, 88 
Liben Formation, Bohemia 15 
Lingula 19 

attenuata 31-2 

plumbea 38 
Lingulacea 25-42 
Lingulella 12, 18, 28, 31 

decorticata 31 

displosa 10, 19-20, 28-9 ; pi. 2, figs. 2-8 
petila 10, 18, 29-31 ; pi. 2, figs. 9-11 ; 
pl- 3- fig- 1 

lingulaeformis 31 

rideanensis 31 
Lingulellinae 28-35 
Lingulida 25-42 
Linoporellidae 114-5 
Llandeilo 12-3, 15*, 16 

Flags 13 
Llandovery 16 
Llanvirn 11 -2, 15*, 19 
Longvillian 14 



Macrocoelia llandeiloensis 13 
Mcewanella 78-81 

berwynensis 79 

sp. 10, 23, 78-9 ; pl. 13, figs. 3, 5 
Meadowtown Beds 9-11, 12*, 13-4, 15*, 

19, 20, 21 
Medesia 68-9 
Monobolina 18-9, 38 

plumbea 10, 12, 17, 38-40 ; pl. 5, figs. 
2-10 ; pl. 6, fig. 1 
Morocco 8, 11, 13-4 



Murinella 141 

partita 141 

sp. 10, 20, 141 ; pl. 25, figs. 2-3 
Mytton Flags 9- 11, 12*, 15*, 16, 17, 18 

Nant Hir Mudstones 14 
Nicolella 14, 22, 57 

cf. actoniae 10, 14, 22-3, 57-60 ; pl. 9, 
figs. 1-6 

actoniae obesa 58-9 
Nocturniella 88 

nocturna 89 

sp. 10-1, 17, 88-9 ; pl. 14, figs. 17, 20 

Obolidae 11, 25-38 
Obolinae 25-8 
Obolus 18, 25 

? biconvex a 26 

? nitens 26 

subditivus 10, 18, 25-6 ; pl. 1, figs. 1-8 

sp. 10, 17, 26 ; pl. 1, figs. 9-10 
Onniella 15, 21-2, 98 

ostentata 98-9 

lepida 10, 14, 22-3, 98-102; pl. 16, 
figs. 9-14 

soudleyensis 99 

sp. 10, 21, 99 ; pl. 16, figs. 15, 18 
Onychoplecia 126 
Orbiculoidea 47-8 

sp. 10, 47 ; pl. 7, fig. 1 
Orbiculoideinae 47-8 
Ordovician 1-158 
Orthacea, orthaceans 7, 49-86 
Orthambonites 22, 24*, 53 

bellus 54 

calligramma 53, 55 

exopunctata 10, 23, 53-7 ; pl. 8, figs. 
10-7 

friendsvillensis 54 

orbicularis 53 

playfairi 54 

cf. rotundiformis 55 

sp. 10, 22 
Orthida 49-126 
Orthidae 50-63, 68 
Orthidina 49-115 
Orthinae 50-7 
Orthis 18, 50 

actoniae 57 

cf. callactis 10-1, 17, 50-1 ; pl. 7, figs. 
10-1, 13-6 

calligramma 53 
proava 52 



INDEX 



161 



Carausii 52 

panderiana 53 

partita 74 

sp. 10, 18, 51 ; pi. 8, fig. 1 
Oslogonites II, 115 

costellatus 115 

? sp. 10, 17, 115 ; pi. 19, figs. 12-3 
Oxoplecia 124, 126 

costellata 125 

dorsata 126 

gibbosa 123 

mutabilis 123 

nantensis 125-6 

cf. nantensis 10, 13, 20, 125-6 ; pi. 20, 
figs. 9-13 

pennsylvanica 123 

sibirica 125 



palaeoecology 16, 23 
palaeogeography 8 
Palaeoglossa 12, 18-9, 21, 31-2 

attenuata 10, 19-23, 32-3, 34-5 ; pi. 3, 
figs. 2-13 

myttonensis 10, 16-7, 33-5 ; pi. 4, 

figs- i-5 
Palaeostrophomena 14, 22, 126 
magnifica 127 

sp. 10, 23, 126-7; pi. 21, figs. 7, 
10-1, 13 
Parastrophinella 151 
costata 152 
musculosa 10, 21, 151-2 ; pi. 28, figs. 

9-13. 17 

sp. 10, 22, 152-3 ; pi. 28, figs. 14-5 
Parastrophinidae 151 -3 
Paterula 11, 18-9, 21, 40 

cf. bohemica 10, 17-20, 40-1, 42 ; pi. 6, 
figs. 2- 1 1 

cf. perfecta 10, 21-2, 41-2; pi. 6, figs. 
12-3 
Paterulidae 40-2 
Paurorthis 108 
Pembrokeshire 14 
Pentamerida 150-3 
Petrocrania 48 

dubia 10, 22-3, 48-9 ; pi. 7, figs. 5-6, 8 

inexpectata 49 
Platymena 143, 146 
Platystrophia 14, 22, 68, 76, 79-81 

caelata 10, 22-3, 76-7, 78 ; pi. 12, figs. 
13-4, 16-9 

chama 77 

dentata 78 



cf. major 10, 14, 21-2, 78 ; pi. 13, figs. 
1-2, 4 

precedens 78 
major 78 
Platystrophiinae 68, 76-81 
Plectambonitacea, plectambonitaceans 7, 

126-40 
Plectorthidae 68, 71-81 
Plectorthinae 71-6 
Plectorthis 71 

whitteryensis 10, 22-3, 71-3; pi. 11, 
figs. 10, 13, 16-7 ; pi. 12, fig. 1 

sp. 10, 22, 73 ; pi. 11, figs. 12, 15 
Plesiomys (Retrosistria) 61 
Porambonitacea 68, 150-3 
Portrane Limestone 14 
Productorthinae 57-60 
Protoskenidioides 7, 18, 83-5, 84* 

revelata 10, 16-7, 83, 85-6 ; pi. 14, figs. 

4-!3 
Pseudobolus 31 

Pseudolingula 12, 18-9, 35, 36* 
quadrata 38 
spatula 10, 17-20, 36-8 ; pi. 4, figs. 

6-14 ; pi. 5, fig. 1 
sp. 11 

Rafinesquina 19, 21, 141-3 

delicata 10, 13, 19-20, 141-3, 144-5 ; 

pi. 25, figs. 6-13 
expansa 143 
simplex 143 

sp. 10, 20, 22-3, 143 ; pi. 25, fig. 1 ; 
pi. 26, fig. 1 
Rafinesquininae 141 -5 
Resserella immatura 109 
Retrosistria 61 
Reuschella 14-5, 22, 104 
horderleyensis 105 

carinata 10, 22-3, 104-6 ; pi. 17, 

figs. 4-9 
undulata 105 
Rhynchonellacea 153-4 
Rhynchonellida 153-4 
Rorrington Beds 9-11, 12*, 13, 15*, 16, 19, 

20 
Rostricellula 153 
ambigua 154 
? minuta 154 
sparsa 10, 21, 153-4 ; pi. 27, figs. 3-4, 7 

Salacorthis 7, 14, 79-80, 81* 
costellata 10, 21, 79, 80-1 ; pi. 13, 

figs. 6-13 



I"-' 



INDEX 



Salopia 14, 114 
salteri 114-5 

cf. salteri 10, 14, 21, 114 ; pi. 19, figs. 10-1 
sp. io, 22-3, 114-5 ; pi. 19, figs. 6-9 
Sarka Formation, Bohemia II, 42, 152 
Schizomania 12, 18-9, 21, 44 
filosa 46 

salopiensis 10, 18-21, 44-6 ; pi. 6, 
figs. 22-6 
Schizophoriidae 87-9 
Schizotreta 47 
elliptica 48 
microthyris 48 

transversa 10, 20, 47-8 ; pi. 7, figs. 2-3, 7 
sp. 10, 17-9, 48 ; pi. 7, fig. 4 
Schmidtites 18-9, 21, 26 

? simplex 11, 20, 22, 26-7, 28 ; pi. 1, 
figs. 1 1 -5 

subcircularis 11, 17-8, 27-8; pi. 1, 
figs. 16-7 ; pi. 2, fig. 1 
Scotland 11 
Sericoidea 22, 139-40 
abdita 140 
cf. abdita 11, 22, 139-40; pi. 24, figs. 

8-9, 12, 15, 17-8 ; pi. 28, fig. 16 
aff. abdita 140 
Shelve District, Shropshire 1-158 
Shropshire 1-158 
Skenidiidae 82-6 
Skenidioides 13, 82-3, 85 

cf. costatus n, 14, 20-1, 23, 82-3, 85 ; 
pi. 13, figs. 14-6 ; pi. 14, figs. 1-3 
Soudleyan 14, 15* 
Sowerbyella 13-4, 21-2, 130, 136 

antiqua 13, 130-1, 132, 135-9 ; pi. 22, 

figs. 4, 7-14 ; pi. 23, figs. 1, 3-4 
cf. antiqua 11, 20, 131-2, 137-9; pi. 23, 

figs. 2, 5-13 
multiseptata 11, 21, 132-3, 135-9; pi. 

23, figs. 14-9 ; pi. 24, figs. 1, 3 
sericea 136 

cf. sericea II, 14, 22-3, 134-9 ; pi. 24, 
figs. 11, 13-4, 16 

permixta 11, 14, 22, 133-4, 136-9 ; pi. 
24, figs. 2, 4-7, 10 
sp. 11, 20 
Sowerbyellidae 130-40 
Sowerbyellinae 130-9 
Spiriferida 154 
Spy Wood Grit 7, 9-11, 12*, 13-4, 15*, 

16, 21 
Stapeley Shales, Volcanic Group 9- 11, 12*, 

15*. 18 
statistical tests 23-4 



Stiperstones Quartzite II, 15* 

Strachan, Dr I. 8, 12 

Strophomenacea, strophomenaceans 7, 140- 

5° 
Strophomenida 126-50 
Strophomenidae 140-8 
Strophomenidina 126-50 
subduction zones 8 
Syntrophiidae 150- 1 
Syntrophiidina 150-3 
systematic methods 23-5 



Tachilla Shales, Morocco 11 
Tazzarinia 13, 75 

drotae 76 

elongata 11, 20, 75-6; pi. 12, figs. 
9-12, 15 

foraminosa 76 
Tissintia 13, 19, 107-8 

immatura 11, 13, 20, 109-14; pi. 18, 
figs. 10, 12-5 ; pi. 19, figs. 1-5 
plana 13 

prototypa 11 -3, 19, 107, 108-9, 11 1-4 ; 
pi. 17, figs. 15-9 ; pi. 18, figs. 1-9, 11 
Trematidae 44-6 
Trigonirhynchiidae 153-4 
trilobites 8, 13, 16 
Triplesia 116 

deformata 117 

simplex 117 

subcarinata 117 

sp. 11, 23, 116-7 ; pi. 19, fig. 16 
Triplesiacea, triplesiaceans 7, 116-26 
Triplesiidae 116-26 
Triplesiidina 116-26 



Viruella 135 



Wales 7, 11 

successions 12-3 
Weston Beds 9-12, 12*, 15*, 19 
Whittard, Professor W. F. 8, 62, 154 
Whittardia 7, 22, 60-1, 61*, 62 

paradoxica 11, 23, 62-3; pi. 9, figs. 
7-15 ; pi. 10, figs. 1-2, 4 
Whittardiinae 7, 60-1, 62-3 
Whittery Shales 7, 9-11, 12*, 14, 15*, 

21-2, 23 

Volcanic Group 15*, 21 



INDEX 163 

Wright, Dr A. D. 155 Zahorany Formation, Bohemia 15, 49 

Zygospira 154 
Xenelasma 150 sp. 11, 21, 154 ; pi. 27, figs. 12-3 

Xenelasmatinae 150-1 Zygospirinae 154 



Professor Alwyn Williams, Ph.D., F.G.S., M.R.I.A., F.R.S.E., F.R.S. 

Department of Geology 

The University of Birmingham 

P.O. Box 363 

Birmingham B15 2TT 

Accepted for publication 1 February 1974 



PLATE i 
Obolus subditivus sp. nov. (p. 25) 

Hope Shales, near Brithdir Farm, Shropshire. 

Figs, i, 3, 2. Holotype, BB 35573a, b. Latex casts of the exterior and interior and the 
internal mould of a pedicle valve, x 5-6. 

Figs. 5, 6. Paratype, BB 35574a, b. Latex cast of the exterior and the internal mould of 
a pedicle valve, X 6-3, X5-8. 

Figs. 4, 8, 7. Paratype, BB 35575a, b. Latex casts of the exterior and interior and the 
internal mould of a brachial valve, X 5-5. 

Obolus sp. (p. 26) 

Mytton Flags, Bergam Quarry, Shropshire. 
Fig. 9. BB 35581a. Exfoliated pedicle valve, X 4-3. 

Mytton Flags, near Wood House, Shropshire. 
Fig. 10. BB 35582. Exfoliated brachial valve, X4"5- 

Schmidtites ? simplex sp. nov. (p. 26) 

Rorrington Beds, Holywell Brook, Shropshire. 
Fig. 11. Holotype, BB 37134a. Internal mould of a pedicle valve, x 8. 
Fig. 14. Paratype, BB 37138. Internal mould of a brachial valve, x 6-7. 

Rorrington Beds, Deadman's Dingle, Shropshire. 
Fig. 12. Paratype, BB 37139. Internal mould of a pedicle valve, x 8. 

Rorrington Beds, near the Mount, Shropshire. 
Fig. 13. Paratype, BB 37136. Exterior of a pedicle valve, xg. 

Rorrington Beds, near Desert, Shropshire. 
Fig. 15. Paratype, BB 37140a. Internal mould of a pedicle valve, x8-3. 

Schmidtites ? simplex subcircularis sp. et subsp. nov. (p. 27) 

Stapeley Shales, Whitsburn Dingle, Shropshire. 
Fig. 16. Holotype, BB 37141. Exterior of a pedicle valve, X9-2. 
Fig. 17. Paratype, BB 37142b. Latex cast of the external mould of a pedicle valve, x 8. 



Bull. By. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE i 




PLATE 2 
Schmidtites ? simplex subcircularis sp. et subsp. nov. (p. 27) 

Stapeley Shales, Whitsburn Dingle, Shropshire. 
Fig. 1. Paratype, BB 37143a. Internal mould of a brachial valve, x 6. 

Lingulella displosa sp. nov. (p. 28) 

Meadowtown Beds, Waitchley, Shropshire. 
Fig. 2. Paratype, BB 37146a. Internal mould of a pedicle valve, X3-8. 

Meadowtown Beds, near Meadowtown Quarry, Shropshire. 
Figs. 4, 3, 8. Holotype, BB 37144a, b. Latex cast, with detail, of the external mould, and 
the internal mould of a pedicle valve, X 4, X 12, x 3. 

Rorrington Beds, Lower Wood Brook, Shropshire. 
Fig. 5. Paratype, BB 37149. Internal mould of a brachial valve, x 2-8. 

Meadowtown Beds, Little Weston, Shropshire. 
Fig. 6. Paratype, BB 37148a. Internal mould of a pedicle valve, X3T. 
Meadowtown Beds, near Meadowtown Quarry, Shropshire. 
Fig. 7. Paratype, BB 37147a. Internal mould of a brachial valve, X3. 

Lingulella displosa petila sp. et subsp. nov. (p. 29) 

Stapeley Volcanic Group, Whitsburn Dingle, Shropshire. 
Fig. 9. Paratype, BB 37152a. Internal mould of a pedicle valve, X4. 
Fig. 10. Paratype, BB 37151a. Internal mould of a brachial valve, X4. 
Fig. 11. Holotype, BB 37150b. Latex cast of the external mould of a pedicle valve, x 9. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 2 




PLATE 3 
Lingulella displosa petila sp. et subsp. nov. (p. 29) 

Stapeley Volcanic Group, Whitsburn Dingle, Shropshire. 
Fig. 1. Holotype, BB 37150a. Internal mould of a pedicle valve, x 4-5. 

Palaeoglossa attenuata (J. de C. Sowerby) (p. 32) 

Meadowtown Beds, Lower Wood Brook, Shropshire. 
Fig. 2. BB 37109a. Internal mould of a brachial valve, x 3-4. 

Meadowtown Beds, Little Weston, Shropshire. 
Fig. 3. BB 37110a. Internal mould of a pedicle valve, x 5. 

Betton Beds, near junction of Holywell and Whitehouse Brooks, Shropshire. 
Figs. 4, 11. BB 37117a, b. Internal moulds of pedicle and brachial valves, X 6-5, x 6-9. 
Fig. 12. BB 37116a. Internal mould of a pedicle valve, X47. 

Meadowtown Beds, near Meadowtown Quarry, Shropshire. 
5. BB 37113. Exterior, with detail, of a pedicle valve, X6-5, x 13. 
BB 37115. Internal mould of a pedicle valve, x 2-8. 

Meadowtown Beds, near Meadowtown Chapel, Shropshire. 
BB 37111a. Internal mould of a ventral pseudointerarea, x 17-3. 
BB 371 12. Latex cast of the exterior of a pedicle valve, x 4-6. 
BB 37114a. Internal mould of a brachial valve, x 5-4. 

Rorrington Beds, in stream near Meadowtown Chapel, Shropshire. 
Fig. 13. BB 37118b. Internal mould of a brachial valve, x 3-5. 



Figs 
Fig. 


'• 7. 
8. 


Fig. 


6. 


Fig. 
Fig. 


9- 
10. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. u 



PLATE 3 




m -■'■$ vS'M»: 



PLATE 4 
Palaeoglossu myttonensis sp. nov. (p. 33) 

Mytton Flags, Whitegrit School, Shropshire. 

Figs. 2, 1. Paratype, BB 37122a, b. External and internal moulds of disarticulated valves, 
X4-9, X5-8. 

Fig. 4. Holotype, BB 371 19. Internal mould of a pedicle valve and an external mould of 
the complementary brachial valve, x 4-9. 

Mytton Flags, near Blakemoorflat, Shropshire. 
Fig. 3. Paratype, BB 37123. Internal mould of a brachial valve, x 5-2. 

Mytton Flags, New Perkin's Level, Shropshire. 
Fig. 5. Paratype, BB 37121a. Internal mould of a brachial valve, x 6. 

Pseudolingula spatula sp. nov. (p. 36) 

Weston Beds, near Lyde, Shropshire. 
Fig. 6. Paratype, BB 37126. Internal mould of a pedicle valve, x 2-9. 

Figs. 7, 10. Paratype, BB 37128a, b. Latex cast of the exterior and the internal mould of a 
brachial valve, x 3, X 3-2. 

Fig. 9. Paratype, BB 37125a. Internal mould of a brachial valve, X3. 

Weston Beds, Cwm Dingle, Shropshire. 
Fig. 8. Holotype, BB 37124a. Internal mould of a brachial valve, x 3. 

Rorrington Beds, Deadman's Dingle, Shropshire. 
Fig. 11. BB 37130a. Internal mould of a pedicle valve, X5"5. 

Meadowtown Beds, Minicop Farm, Shropshire. 
Figs. 12, 13. BB 37129. Fragments of the external lateral area and the interior of a pedicle 
valve, x 6, x 6-4. 

Mytton Flags, River Camlad, Shropshire. 
Fig. 14. BB 35583. Exfoliated pedicle valve, x 3. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 4 




PLATE 5 
Pseudolingula spatula sp. nov. (p. 36) 

Weston Beds, near Lyde, Shropshire. 
Fig. 1. Paratype, BB 37128b. Detail of the latex cast of the external mould of a brachial 
valve, X21. 

Monobolina plumbea (Salter) (p. 38) 

Mytton Flags, Old Perkin's Level, Shropshire. 
Fig. 2. BB 35467. Latex cast of the interior of a pedicle valve, x 3-4. 

Mytton Flags, near Snailbeach Coppice, Shropshire. 
Figs. 3, 10. BB 35470a, b. Latex cast of the exterior and the internal mould of a brachial 
valve, X5'4, X3*6. 

Mytton Flags, New Perkin's Level, Shropshire. 
Figs. 4, 7. BB 35468. Latex cast, with detail, of the exterior of a pedicle valve, X 6, x 12. 

Mytton Flags, Perkin's Beach, Shropshire. 
Figs. 5, 6. BB 35473a. Latex cast of the interior and the internal mould of a pedicle 
valve, x 2-6. 

Mytton Flags, near St Luke's Church, Shropshire. 
Fig. 8. BB 35472. Internal mould of a pedicle valve, x 2-4. 

Mytton Flags, Mytton Batch, Shropshire. 
Fig. 9. BB 35469a. Internal mould of a brachial valve, x 2-8. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 5 




PLATE 6 
Monobolina plumbea (Salter) (p. 38) 

Mytton Flags, near Blakemoorflat, Shropshire. 
Fig. 1. BB 35471. Latex cast of the internal mould of a pedicle valve, x 2-6. 

Paterula cf. bohemica Barrande (p. 40) 

Hope Shales, Hope Dingle, Shropshire. 

Figs. 2, 3. BB 35588a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, xii. 

Figs. 5, 4. BB 35589a, b. Latex cast of the exterior and the internal mould of a brachial 
valve, x 11-9. 

Meadowtown Beds, Minicop Farm, Shropshire. 

Fig. 6. BB 35590. Exterior of a pedicle valve, x 6. 

Fig. 7. BB 35591. Exterior of a brachial valve, xn-i. 

Fig. 8. BB 35592. Exterior of a pedicle valve, xn-g. 

Fig. 9. BB 35593. Exterior of a pedicle valve, xn-g. 

Fig. 10. BB 35594. Exterior of a brachial valve, x 12-1. 

Fig. 11. BB 35595. Exterior of a brachial valve, x 10. 

Paterula cf. perfecta Cooper (p. 41) 

Spy Wood Grit, Spy Wood Dingle, Shropshire. 
Fig. 12. BB 35597. Exterior of a brachial valve, x 11. 
Fig. 13. BB 35596. Exterior of a pedicle valve, xii-5- 

Apsotreta sp. (p. 42) 

Mytton Flags, near Wood House, Shropshire. 
Figs. 14, 15. BB 35565a. Ventral and lateral views of the internal mould of a pedicle valve, 
X7. X5. 

Conotreta stapeleyensis sp. nov. (p. 43) 

Stapeley Shales, Leigh Hall, Shropshire. 

Fig. 16. Holotype, BB 35566. Internal mould of a pedicle valve, x 7. 

Fig. 17. Paratype, BB 35568. Internal mould of a brachial valve, X57. 

Figs. 18, 19. Paratype, BB 35569. Latex cast of the interior and the internal mould of a 
brachial valve, x 7, x 6. 

Fig. 20. Paratype, BB 35570. Internal mould of a brachial valve with adherent shell 
posteriorly, x 6. 

Fig. 21. Paratype, BB 35571b. Latex cast of the exterior of a pedicle valve, X 5-7. 

Schizocrania salopiensis sp. nov. (p. 44) 

Rorrington Beds, near Desert, Shropshire. 
Fig. 22. Paratype, BB 35361. Exterior of a brachial valve, x 7-1. 
Fig. 23. Paratype, BB 35362a. Internal mould of a brachial valve, X 4. 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 24. Paratype, BB 35359. Exterior of a brachial valve, x 3-5. 
Fig. 25. Holotype, BB 35358a. Internal mould of a brachial valve, x 2-2. 

Betton Beds, near Little Weston, Shropshire. 
Fig. 26. Paratype, BB 35404a. Internal mould of a brachial valve, x 7. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 6 




mtesMHk 26 



PLATE 7 
Orbiculoidea sp. (p. 47) 

Stapeley Shales, Shropshire (unlocated exposure). 
Fig. 1. BB 35576. Latex cast of the exterior of a brachial valve, x 5. 

Schizotreta transversa sp. nov. (p. 47) 

Meadowtown Beds, near Meadowtown Quarry, Shropshire. 
Fig. 2. Paratype, BB 35578. Exfoliated pedicle valve, x6-2. 
Fig. 7. Holotype, BB 35577. Exfoliated pedicle valve, X67. 

Meadowtown Beds, Middleton, Shropshire. 
Fig. 3. Paratype, BB 35579. Exfoliated brachial valve, xyi. 

Schizotreta sp. (p. 48) 

Mytton Flags, New Perkin's Level, Shropshire. 
Fig. 4. BB 35580. Latex cast of the exterior of a pedicle valve, x 8. 

Petrocrania dubia sp. nov. (p. 48) 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Figs. 8, 6, 5. Holotvpe, BB 35405a, b. Latex casts of the exterior and interior and the 
internal mould of a brachial valve, x 2-6, x 3-5, X 3. 

Hesperonomia sp. (p. 49) 

Mytton Flags, Woodside House, Shropshire. 
Figs. 12, 9. BB 35334a, b. Latex casts of the external and internal moulds of a brachial 
valve, x8-3, X5. 

Orthis cf. callactis Dalman (p. 50) 

Mytton Flags, Crowsnest Dingle, Shropshire. 
Figs. 10, 11. BB 35498a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x 3-6. 

Mytton Flags, near St Luke's Church, Shropshire. 
Fig. 13. BB 35500. Latex cast of the exterior of a brachial valve, x 8-4. 
Figs. 15, 14. BB 35497. Latex casts showing dorsal and ventral views of conjoined valves, 

X3-4- 

Fig. 16. BB 35499. View of a fragment of the internal mould of a brachial valve and the 
external mould of the complementary pedicle valve, X 12-5. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 7 




PLATE 8 
Orthis sp. (p. 51) 

Stapeley Volcanic Group, Leigh, Shropshire. 
Fig. 1. BB 35425. Latex cast of the interior of a brachial valve, X6-5. 

Lenorthis cf. proava (Salter) (p. 52) 

Mytton Flags, Snailbeach, Shropshire. 

Figs. 3, 2. BB 35501a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x 6-6, X 3-4. 

Figs. 5, 4. BB 35502a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, X3'5. 

Mytton Flags, Perkin's Beach, Shropshire. 
Figs, 6, 9. BB 35504a. Latex cast and mould of the interior of a brachial valve, x 3-7. 
Figs, 7, 8. BB 35503a, b. Latex cast of the exterior and interior of a brachial valve, x 3-5. 

Orthambonites exopunctata sp. nov. (p. 53) 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Figs. 10, 11. Holotype, BB 35371. General and detailed views of the exterior of a brachial 
valve, x 3-3, X 10. 

Fig. 12. Paratype, BB 35372. Exterior of a brachial valve, X3-2. 

Figs. 13, 16. Paratype, BB 35376a, b. Interior and internal mould of pedicle valve, x 3-1, 

X3-9- 
Fig. 14. Paratype, BB 35373. Latex cast of the interior of a brachial valve, x 3-5. 

Whittery Shales, Whittery Quarry, Shropshire. 
Fig. 15. Paratype, BB 35375. Internal mould of a pedicle valve, X4-2. 

Whittery Shales, River Camlad, Shropshire. 
Fig. 17. Paratype, BB 35374a. Latex cast of the interior of a brachial valve, X4 - 7- 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 8 




PLATE g 
Nicolella cf. actoniae (J. de C. Sowerby) (p. 57) 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs, i, 4. BB 35327a, b. Latex casts of the exterior and interior of a brachial valve, x 6. 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 2. BB 35330. Latex cast of the interior of a pedicle valve, x 3. 

Whittery Shales, River Camlad, Shropshire. 
Fig. 3. BB 35328a. Latex cast of the interior of a brachial valve, x 3-2. 

Whittery Shales, near Marrington Farm, Shropshire. 
Fig. 5. BB 35331. Internal mould of a pedicle valve, x 2-8. 
Fig. 6. BB 35329. Latex cast of the exterior of a brachial valve, x 3-2. 

Whittardia paradoxica gen. et sp. nov. (p. 62) 

Whittery Shales, Whittery Quarry, Shropshire. 

Figs, 7, 8. Paratype, BB 35385. Latex cast and mould of the interior of a brachial valve, 
X3-6, Xf2. 

Figs. 9, 13. Holotype, BB 35381b. Latex cast, with detail, of the exterior of a brachial 
valve, x 5-2, x 13. 

Fig. 10. Paratype, BB 35403. Internal mould of a pedicle valve, x 5. 

Figs. 12, n. Paratype, BB 35383b. Latex cast, with detail, of the exterior of a pedicle 
valve, X 5, X 12-5. 

Fig. 14. Paratype, BB 35382a. Internal mould of a pedicle valve, x 2-6. 

Fig. 15. Paratype, BB 35402a. Latex cast of the interior of a brachial valve, X4'3. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 9 




PLATE 10 
Whittardia paradoxica gen. et sp. nov. (p. 62) 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 1. Paratype, BB 35384b. Latex cast of the exterior of a brachial valve, x 4-8. 

Whittery Shales, Whittery Quarry, Shropshire. 
Fig. 2. Paratype, BB 35383a. Internal mould of a pedicle valve, X5'i. 
Fig. 4. Paratype, BB 35402b. Latex cast of the exterior of a brachial valve, x 5-7. 

Dolerorthis cf. tenuicostata Williams (p. 63) 

Whittery Shales, Whittery Quarry, Shropshire. 

Figs. 3, 5. BB 35460a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x6-i, X3-6. 

Figs. 7, 10, 13. BB 35459a, b. Latex casts of the exterior and the interior and the internal 
mould of a brachial valve, X 3-3, X3-6, X 3-6. 

Glyptorthis viriosa sp. nov. (p. 64) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 6, 14. Holotype, BB 35505a, b. Latex casts of the exterior and interior of a brachial 
valve, X5. 

Fig. 8. Paratype, BB 35509. Internal mould of a pedicle valve, X4*8. 

Fig. 9. Paratype, BB 35506a. Internal mould of a pedicle valve, x 5-3. 

Fig. 11. Paratype, BB 35508a. Latex cast of the interior of a brachial valve, X5. 

Fig. 12. Paratype, BB 35507. Internal mould of a pedicle valve, X5. 

Meadowtown Beds, Minicop Farm, Shropshire. 
Fig. 15. BB 35520b. Latex cast of the exterior of a brachial valve, x 7-5. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 10 




PLATE ii 
Glyptorthis viriosa sp. nov. (p. 64) 

Meadowtown Beds, Quinton's Quarry, Shropshire. 
Figs, i, 2. BB 35521a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x6-8, X5'i. 

Meadowtown Beds, Minicop Farm, Shropshire. 
Fig. 4. BB 35520a. Internal mould of a brachial valve, X5~5. 

Diparelasma sp. (p. 71) 

Mytton Flags, near Mytton's Beach, Shropshire. 
Fig. 3. BB 35335a, b. Latex casts of the interior of a brachial valve and the exterior of the 
complementary pedicle valve, x 8. 

Mytton Flags, near Blakemoorflat, Shropshire. 
Figs. 7, 8. BB 35336a, b. Latex casts of the exterior and interior of a brachial valve, x 5. 

Astraborthis uniplicata gen. et sp. nov. (p. 70) 

Mytton Flags, near Snailbeach Reservoir, Shropshire. 

Figs. 6, 5, 11. Paratype, BB 35325a, b. Latex casts of the exterior and interior and the 
internal mould of a brachial valve, x 4-7, X 4, X 4. 

Figs. 14, 9. Holotype, BB 35324a, b. Latex cast of part of the exterior and the internal 
mould of a pedicle valve, X 4, X 2-8. 

Plectorthis sp. (p. 73) 

Aldress Shales, Ox Wood Dingle, Shropshire. 
Figs. 15, 12. BB 35466a. Latex cast and mould of the interior of a pedicle valve, x 2-7. 

Plectorthis whitteryensis sp. nov. (p. 71) 

Whittery Shales, River Camlad, Shropshire. 
Figs. 13, 10. Holotype, BB 35462a, b. Latex cast of the exterior and the internal mould of 
a pedicle valve, X 3-4. 

Fig. 16. Paratype, BB 37158a. Internal mould of a pedicle valve, X3-5. 
Fig. 17. Paratype, BB 35463a. Internal mould of a brachial valve, X3'5- 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 

m 



PLATE n 




PLATE 12 
Plectorthis whitteryensis sp. nov. (p. 71) 

Whittery Shales, River Camlad, Shropshire. 
Fig. 1. Paratype, BB 35464a. Latex cast of the interior of a brachial valve, X3"5- 

Desmorthis ? sp. nov. (p. 73) 

Stapeley Volcanic Group, Perkin's Beach, Shropshire. 
Fig. 2. BB 35339a. Internal mould of pedicle valve, X5. 
Fig. 3. BB 35338a. Internal mould of pedicle valve, X 6-3. 

Mytton Flags, near Blakemoorflat, Shropshire. 
Figs. 8, 4. BB 35337a, b. Latex casts of the exterior and interior of a brachial valve, 
X4. X3. 

Gelidorthis cf. partita (Barrande) (p. 74) 

Rorrington Beds, Spy Wood Brook, Shropshire. 
Fig. 5. BB 35419a. Internal mould of a brachial valve, x 7. 
Fig. 6. BB 35418a. Internal mould of a pedicle valve, x 5. 
Fig. 7. BB 37159a. Internal mould of a pedicle valve, X5. 

Tazzarinia elongata sp. nov. (p. 75) 

Meadowtown Beds, near Waitchley, Shropshire. 

Figs. 9, 11, 10. Holotype, BB 35332a, b. Latex casts of the exterior and interior and the 
internal mould of a brachial valve, x 5-5, X 6-2, x 5. 

Figs. 12, 15. Paratype, BB 35333a, b. Latex cast of the exterior and the internal mould of 
a pedicle valve, x6-i. 

Platystrophia caelata sp. nov. (p. 76) 

Whittery Shales, River Camlad, Shropshire. 
Figs. 16, 13. Paratype, BB 35495a, b. Latex cast of the exterior and the internal mould of 
a pedicle valve, x 4, x 2-5. 

Fig. 17. Paratype, BB 35494a. Internal mould of a pedicle valve, x 2-7. 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs. 19, 14, 18. Holotype, BB 35493a, b. Latex cast and detail of the mould of the exterior 
and the internal mould of a brachial valve, x 2-6, x 14, x 2-6. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 12 




PLATE 13 
Platystrophia cf. major Williams (p. 78) 

Aldress Shales, Ox Wood Dingle, Shropshire. 
Fig. 1. BB 35586a. Internal mould of a brachial valve, x 7. 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 4, 2. BB 35587a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x6-2. 

Mcewanella sp. (p. 78) 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 3. BB 35417. Internal mould of a pedicle valve, X2-i. 
Fig. 5. BB 35416. Exterior of a pedicle valve, x 2-4. 

Salacorthis costellata gen. et sp. nov. (p. 80) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 7, 11, 6, 10. Holotype, BB 37153a, b. Latex casts and moulds of the exterior and in- 
terior of a brachial valve, x 7-6, x 4-9, X 7-6, x 7-6. 

Fig. 8. Paratype, BB 37154. Internal mould of a pedicle valve, X3*i. 

Fig. 12. Paratype, BB 37155. Latex cast of the exterior of a brachial valve, x 3-6. 

Fig. 13. Paratype, BB 37157. Latex cast of the exterior of a brachial valve, x 5-1. 

Spy Wood Grit, Spy Wood Brook, Shropshire. 
Fig. 9. Paratype, BB 37156. Latex cast of the exterior of a brachial valve, x 6. 

Skenidioides cf. costatus Cooper (p. 82) 

Spy Wood Grit, Spy Wood Brook, Shropshire. 
Fig. 14. BB 35413. Latex cast of the exterior of a pedicle valve, x 12. 

Whittery Shales, River Camlad, Shropshire. 
Figs. 16, 15. BB 35409. Latex cast and mould of the interior of a brachial valve, X57, x 7. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. u 



PLATE 13 




PLATE 14 
Skenidioides cf. costatus Cooper (p. 82) 

Whittery Shales, Hockleton Bridge, Shropshire. 
Fig. 1. BB 35410. Internal mould of a pedicle valve, x 7-2. 

Spy Wood Grit, Spy Wood Brook, Shropshire. 
Fig. 2. BB 3541 1. Internal mould of a pedicle valve, x i3 - 3- 
Fig. 3. BB 35412. Internal mould of a pedicle valve, x 12-5. 

Protoskenidioides revelata gen. et sp. nov. (p. 85) 

Mytton Flags, near Gravels, Shropshire. 
Paratype, BB 35395a. Internal mould of a pedicle valve, x 13-6. 
Paratype, BB 35394b. Latex cast of the exterior of a brachial valve, x 10. 
Paratype, BB 35397b. Latex cast of the exterior of a brachial valve, x 12-5. 
6. Paratype, BB 35400a. Latex cast and mould of the interior of a brachial valve, 

Mytton Flags, near Shelve Church, Shropshire. 

Holotype, BB 35387. Internal mould of a brachial valve, x 12-4. 

Paratype, BB 35389. Internal mould of a pedicle valve, x 15. 

Paratype, BB 35399a. Internal mould of a brachial valve, X 14. 
Paratype, BB 35393. Internal mould of a brachial valve, x i4 - 5- 
Paratype, BB 35398a. Internal mould of a brachial valve, x 10. 

Drabovia cf. fascicostata Havlicek (p. 87) 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 14. BB 35319b. Latex cast of the exterior of a pedicle valve, xy-i. 
Figs. 15, 18. BB 35316a, b. Latex casts of the exterior and interior of a brachial valve, x 8. 
Fig. 16. BB 35317a. Internal mould of a brachial valve, x 6-4. 
Fig. 19. BB 35318a. Internal mould of a pedicle valve, X67. 

Nocturniella sp. (p. 88) 

Mytton Flags, Snailbeach, Shropshire. 
Figs. 17, 20. BB 35341. Latex cast and mould of the interior of a brachial valve, x 7. 



Fig. 


4- 


Fig. 


9- 


Fig. 


11. 


Figs 


'<• !2, 


X9, x 14. 


Fig. 


5- 


Fig. 


7- 


Fig. 


8. 


Fig. 


10. 


Fig. 


13- 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 14 




PLATE 15 
Dalmanella parva Williams (p. 89) 

Lower Llandeilo sandstones, near Ysgubor-wen Farm, Llandeilo, Carmarthenshire. 
Fig. 1. BB 35436. Internal mould of a brachial valve, x 8. 
Fig. 2. BB 35437. Latex cast of the exterior of a brachial valve, x 6. 
Figs, 3, 7. BB 35434. Latex cast and mould of the interior of a brachial valve, X 6-5. 
Fig. 4. BB 35435. Internal mould of a pedicle valve, x 6-9. 

Dalmanella salopiensis sp. nov. (p. 90) 

Meadowtown Beds, near Minicop Farm, Shropshire. 
Figs. 8, 5. Holotype, BB 35446a, b. Latex cast of the exterior and the internal mould of a 
brachial valve, x 8, x 6. 

Fig. 11. Paratype, BB 35448a. Internal mould of a brachial valve, X63. 

Meadowtown Beds, Quinton's Quarry, Shropshire. 
Figs. 12, 6. Paratype, BB 35450a, b. Latex casts of the exterior and interior of a brachial 

valve, x 7-2. 

Meadowtown Beds, near Meadowtown Chapel, Shropshire. 
Figs. 10, 9. Paratype, BB 35447a, b. Latex cast of the exterior and the internal mould of a 
pedicle valve, x 6-9, x 7-8. 

Dalmanella salopiensis gregaria sp. et subsp. nov. (p. 91) 

Spy Wood Grit, Spy Wood Brook, Shropshire. 
Figs. 16, 13. Holotype, BB 35443a, b. Latex cast of the exterior and the internal mould 
of a brachial valve, x 6-4, x 7-5. 

Fig. 14. Paratype, BB 35407a. Internal mould of a pedicle valve, x 7-3. 

Fig. 15. Paratype, BB 35444a. Internal mould of a pedicle valve, X9. 

Fig. 20. Paratype, BB 35445a. Latex cast of the interior of a brachial valve, x 10. 

Dalmanella salopiensis transversa sp. et subsp. nov. (p. 92) 

Aldress Shales, Ox Wood Dingle, Shropshire. 

Figs. 21, 19. Holotype, BB 35438a, b. Latex casts of the exterior and interior of a brachial 
valve, x 6-5. 

Figs. 22, 23. Paratype, BB 35441a, b. Latex cast of the exterior and the internal mould of 
a pedicle valve, x 10. 

Fig. 24. Paratype, BB 35439a. Internal mould of a pedicle valve, X 6-9. 

Figs. 18, 17. Paratype, 35440a, b. Latex cast of the exterior and the internal mould of a 
brachial valve, X 10. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. 1 1 



PLATE 15 




PLATE 16 
Dalmanella salopiensis transversa sp. et subsp. nov. (p. 92) 

Aldress Shales, Ox Wood Dingle, Shropshire. 
Fig. 1. Paratype, BB 35442a. Internal mould of a pedicle valve, x 10. 

Dalmanella eletnentaria sp. nov. (p. 97) 

Mytton Flags, near Shelve Church, Shropshire. 
Figs. 2, 3. Holotype, BB 35352a, b. Latex cast of the exterior and the internal mould of a 
brachial valve, x 7-8. 

Fig. 4. Paratype, BB 35357a. Internal mould of a brachial valve, x 8. 
Fig. 7. Paratype, BB 35355a. Internal mould of a brachial valve, x 8-3. 
Fig. 8. Paratype, BB 35356a. Internal mould of a brachial valve, x 8. 

Mytton Flags, near Wood House, Shropshire. 
Figs. 5, 6. Paratype, BB 35353a, b. Latex cast of the exterior and the internal mould of a 
pedicle valve, x 8. 

Onniella ostentata lepida subsp. nov. (p. 98) 

Aldress Shales, Ox Wood Dingle, Shropshire. 

Fig. 9. Paratype, BB 35457a. Internal mould of a brachial valve, x 5-7. 

Figs. 10, 11. Holotype, BB 35456a, b. Latex cast of the exterior and the internal mould of 
a brachial valve, x 7, X 5-6. 

Fig. 12. Paratype, BB 35458. Latex cast of the interior of a brachial valve, x 8. 

Figs. 13, 14. Paratype, BB 35455a, b. Latex cast of the exterior and the internal mould of a 
pedicle valve, x 6-o, X5-8. 

Onniella sp. (p. 99) 

Spy Wood Grit, Shropshire (unlocated exposure). 
Figs. 15, 18. BB 35572a, b. Latex casts of the exterior and interior of a brachial valve, x 6. 

Harknessella cf. subplicata Bancroft (p. 102) 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 16. BB 35427. Internal mould of a brachial valve, x 6-7. 

Horderleyella cf. plicata Bancroft (p. 102) 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 17. BB 35432. Internal mould of a brachial valve, x 6-6. 

Figs. 19, 20. BB 35431a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x 4. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 16 




Fig. i. BB 35433. 



PLATE 17 
Horderleyella cf. plicata Bancroft (p. 102) 

Aldress Shales, Ox Wood Dingle, Shropshire. 
Internal mould of a brachial valve, X4"5. 



Figs. 

X7-5- 



Fig. 



2, 3- 



Horderleyella sp. (p. 104) 

Meadowtown Beds, near Meadowtown Chapel, Shropshire. 
BB 35420a, b. Internal mould and latex cast of the exterior of a pedicle valve, 



Fig. 


5- 


Holotype, 


BB 35513a 


Fig. 


6. 


Paratype, 


35 355^ 


Fig. 


7- 


Paratype, 


BB 35515. 


Fig. 


8. 


Paratype, 


BB 35514a 


Fig. 


9- 


Paratype, 


BB 35518a 



Reuschella horderleyensis carinata subsp. nov. (p. 104) 

Whittery Shales, near Marrington Farm, Shropshire. 
Paratype, BB 35519a. Internal mould of a brachial valve, x 6. 

Whittery Shales, River Camlad, Shropshire. 

Internal mould of a pedicle valve, x 2-1. 
Internal mould of a pedicle valve, x 1-9. 
Latex cast of the exterior of a pedicle valve, x 2-1. 
Internal mould of a brachial valve, x 2-6. 
Latex cast of the interior of a brachial valve, x 2. 

Heterorthis sp. (p. 106) 

Spy Wood Grit, near Rorrington, Shropshire. 

Figs, 10, 14, 12. BB 35424a, b. Latex casts of the exterior and interior and the internal 
mould of a brachial valve, x 8, x 6, x 4-9. 

Figs, ii, 13. BB 35423a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x 6. 

Tissintia prototypa Williams (p. 108) 

Lower Llanvirn ashy shales, near Llwyn Bedw Farm, Llandeilo, Carmarthenshire. 
Fig. 15. BB 35310. Latex cast of the interior of a brachial valve, x 3-2. 
Fig. 16. BB 35305. Latex cast of the interior of a brachial valve, x 3-3. 
Figs. 18, 17. BB 35309. Latex cast and internal mould of a brachial valve, x 2-3. 
Fig. 19. BB 35308. Latex cast of the exterior of a brachial valve, X3 - 5- 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 17 




PLATE 18 
Tissintia prototypa (Williams) (p. 108) 
Lower Llanvirn ashy shales, near Llwyn Bedw Farm, Llandeilo, Carmarthenshire. 
Fig. i. BB 35306. Internal mould of a pedicle valve, x 3. 
Fig. 2. BB 35307. Latex cast of the exterior of a brachial valve, x 3-3. 

Weston Beds, near Lyde, Shropshire. 
Figs, ii, 3. BB 35311a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, X 1 -9, x 2-4. 

Fig. 6. BB 35315a. Latex cast of the interior of a brachial valve, x 2-5. 

Figs. 8, 7. BB 35314a. Latex cast and internal mould of a brachial valve, x 2-2. 

Weston Beds, near Priestweston, Shropshire. 
Figs. 5, 4. BB 35312a. Latex cast and mould of the interior of a pedicle valve, x 2-6. 

Weston Beds, near Little Weston, Shropshire. 
Fig. 9. BB 35313b. Latex cast of the exterior of a brachial valve, x 2. 

Tissintia immatura (Williams) (p. 109) 

Meadowtown Beds, Quinton's Quarry, Shropshire. 
Fig. 10. BB 35479. Internal mould of a pedicle valve, X27. 

Meadowtown Beds, near Meadowtown, Shropshire. 
Fig. 12. BB 35477a. Latex cast of the internal mould of a brachial valve, x 2. 
Fig. 15. BB 35476b. Latex cast of the exterior of a pedicle valve, x 3-8. 

Meadowtown Beds, Minicop Farm, Shropshire. 
Figs. 13, 14. BB 35474a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x 2, X2-5. 



PLATE i 8 




PLATE 19 
Tissintia immatura (Williams) (p. 109) 

Meadowtown Beds, Meadowtown Quarry, Shropshire. 
Figs. 1,2,5. B 35478a, b. Latex casts of the exterior and interior and the internal mould 
ol a hiaci ; -l v , X3-8, X3"5, X2-5. 

Meadowtown Beds, Minicop Farm, Shropshire. 
Figs. 3, 4. BB 35475a, b. Latex cast of the exterior and the internal mould of a brachial 
valve, X2-5. 

Salopia sp. (p. 114) 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs. 9, 6. BB 35428a, b. Latex cast of the exterior and the internal mould of a brachial 
valve, X4'6, X 3-5. 

Fig. 8. BB 35429a. Internal mould of pedicle valve, x 2-9. 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 7. BB 35430a. Internal mould of pedicle valve, X 3. 

Salopia cf. salteri (Davidson) (p. 114) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 10, 11. BB 35426a. Latex cast and mould of the interior of a brachial valve, X4'2, 
X5- 

Oslogonites ? sp. (p. 115) 

Mytton Flags, near Wood House, Shropshire. 
Figs. 12, 13. BB 35326a, b. Latex cast of the exterior and the internal mould of a brachial 
valve, x 10. 

Kullervo sp. (p. 115) 

Meadowtown Beds, near Meadowtown Chapel, Shropshire. 
Figs. 15, 14. BB 35340a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x8, X5'5. 

Triplesia sp. (p. 116) 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 16. BB 35408. Internal mould of a pedicle valve, x 2-5. 

Bicuspina subquadrata sp. nov. (p. 117) 

Whittery Shales, Whittery Quarry, Shropshire. 
Fig. 17. Paratype, BB 35481b. Latex cast of the exterior of a brachial valve, x 2-1. 
Fig. 19. Paratype, BB 35483. Internal mould of a brachial valve, x 1-9. 

Whittery Shales, River Camlad, Shropshire. 
Fig. 18. Holotype, BB 35480a. Internal mould of a pedicle valve, x i-6. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE i< 




PLATE 20 
Bicuspina subquadrata sp. nov. (p. 117) 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs, i, 2. Paratype, BB 35482a, b. Latex cast of the exterior and the internal mould of a 
pedicle valve, X2-8. 

Bicuspina modesta sp. nov. (p. 118) 

Spy Wood Grit, near Rorrington, Shropshire. 

Figs. 3, 4. Paratype, BB 35512a, b. Latex cast of the exterior and the internal mould of a 
brachial valve, X 3-4. 

Figs. 5, 6. Holotype, BB 35510a, b. Latex cast of the exterior and the internal mould of a 
pedicle valve, X 3-2, X 3-5. 

Figs. 7, 8. Paratype, BB 35511a, b. Latex cast of the exterior and the internal mould of a 
brachial valve, X33. 

Oxoplecia cf. nantensis MacGregor (p. 125) 

Meadowtown Beds, near Waitchley, Shropshire. 
Figs. 10, 9. BB 35320a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x 3. 

Fig. 12. BB 35321. Internal mould of a brachial valve, X4 - 5- 

Fig. 13. BB 35323. Latex cast of the interior of a brachial valve, x 5. 

Meadowtown Beds, near Little Weston, Shropshire. 
Fig. 11. BB 35322a. Internal mould of a brachial valve, x 3-5. 

Caeroplecia plicata gen. et sp. nov. (p. 122) 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs. 15, 14. Paratype, BB 35349a, b. Latex cast of the exterior and the internal mould of 
a brachial valve, x 3. 

Fig. 16. Paratype, BB 35351b. Latex cast of the exterior of a brachial valve, X 3-2. 



Bull. By. Mus. nat. Hist. (Geol.) Suppl. u 



PLATE 20 




PLATE 21 
Caeroplecia plicata gen. et sp. nov. (p. 122) 

Whittery Shales, Whittery Quarry, Shropshire. 

Figs. 1,2. Holotype, BB 35342b. Latex cast, with detail, of the exterior of a pedicle valve, 
Xf6, x 14. 

Fig. 4. Paratype, BB 35346a. Internal mould of a pedicle valve, x 3-4. 

Figs. 8, 6. Paratype, BB 35348a, b. Latex casts of the exterior and interior of a brachial 
valve, x 2-6, x 2-1. 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 3. Paratype, BB 35347a. Internal mould of a pedicle valve, x 3. 

Whittery Shales, River Camlad, Shropshire. 
Fig. 5. Paratype, BB 35350a. Internal mould of a brachial valve, x 4-6. 

Palaeostrophomena sp. (p. 126) 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs. 7, 13. BB 35414a, b. Latex casts of the exterior of a pedicle and the complementary 
brachial valve, x 4. 

Whittery Shales, near Marrington Farm, Shropshire. 
Fig. 10. BB 35415a. Internal mould of a pedicle valve, x 4. 

Whittery Shales, River Camlad, Shropshire. 
Fig. 11. BB 35451a. Internal mould of a pedicle valve, X2-5. 

Eocramatia dissimulata gen. et sp. nov. (p. 129) 

Hope Shales, Brithdir Farm, Shropshire. 
Figs. 14, 9. Holotype, BB 35488a. Latex cast and mould of the interior of a pedicle valve, 

X3-5- 
Figs. 12, 15. Paratype, BB 35490a. Latex cast and mould of the interior of a pedicle valve, 

X3-5- X5-5. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 21 




PLATE 22 
Eocramatia dissimulata gen. et sp. nov. (p. 129) 
Hope Shales, Brithdir Farm, Shropshire. 
Paratype, BB 35489a, b. Latex casts of the exterior and interior of a brachial 



Figs, i, 2. 
valve, X5-8. 

Fig. 3. Paratype, BB 35492. 
Fig. 5. Holotype, BB 35488b. 
Fig. 6. Paratype, BB 35491a. 



Latex cast of the exterior of a brachial valve, x 3-4. 
Latex cast of the exterior of a pedicle valve, X 5-3. 
Internal mould of a brachial valve, x 8. 



Figs. 7, 


4- BB355 


x 7-1. 




Fig. 8. 


BB 35524. 


Fig. 9. 


BB 35530. 


Fig. 10. 


BB 35529. 


Fig. 11. 


BB 35526. 


Fig. 12. 


BB 35532. 


Fig. 13. 


BB 35528. 


Fig. 14. 


BB 35533. 



Sowerbyella antiqua Jones (p. 130) 
Ffairfach Group, near Llangadog, Carmarthenshire. 
25. Latex cast and mould of the interior of a brachial valve, xn-2, 

Latex cast of the exterior of a brachial valve, x 6-2. 
Latex cast of the interior of a brachial valve, x 4-7. 

Latex cast of the interior of a brachial valve, x 8-7. 

Latex cast of the exterior of a brachial valve, X 5-5. 

Latex cast of the interior of a brachial valve, x 6. 

Latex cast of the exterior of a pedicle valve, x 5-5. 

Internal mould of a pedicle valve, x 6. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 22 




PLATE 23 
Sowerbyella antiqua Jones (p. 130) 

Ffairfach Group, near Llangadog, Carmarthenshire. 
Fig. 1. BB 35527. Internal mould of a pedicle valve, X47. 
Fig. 3. BB 35531. Internal mould of a pedicle valve, x 6. 
Fig. 4. BB 35534. Latex cast of the interior of a brachial valve, x 6-5. 

Sowerbyella cf. antiqua Jones (p. 131) 

Meadowtown Beds, near Meadowtown, Shropshire. 
Fig. 2. BB 35541a. Latex cast of the interior of a brachial valve, x 6. 
Fig. 5. BB 35536. Latex cast of the exterior of a brachial valve, x 8-7. 
Fig. 6. BB 35539. Internal mould of a pedicle valve, x 6. 
Fig. 7. BB 35540. Internal mould of a pedicle valve, x 5-8. 

Figs. 8, 9. BB 35537a, b. Internal mould and latex cast of the exterior of a pedicle valve, 
X7-3, X7-9- 
Figs. 12, 10. BB 35542. Latex cast and mould of the interior of a brachial valve, x8. 
Fig. 11. BB 35535. Latex cast of the exterior of a brachial valve, x 6. 
Fig. 13. BB 35538. Internal mould of a brachial valve, x 7. 

Sowerbyella multiseptata sp. nov. (p. 132) 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 14. Paratype, BB 35548b. Latex cast of the exterior of a pedicle valve, x 4-5. 
Figs. 16, 15. Holotype, BB 35544. Latex cast and mould of the interior of a brachial valve, 
X6-2. 
Fig. 17. Paratype, BB 35547a. Latex cast of the interior of a brachial valve, x 6. 
Fig. 18. Paratype, BB 35546a. Internal mould of a pedicle valve, x 7. 
Fig. 19. Paratype, BB 35545b. Latex cast of the exterior of a pedicle valve, x 10. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 23 




PLATE 24 
Sowerbyella multiseptata sp. nov. (p. 132) 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 1. Paratype, BB 35548a. Internal mould of a pedicle valve, x 6-6. 
Fig. 3. Paratype, BB 35545a. Internal mould of a pedicle valve, X 5-8. 

Sowerbyella cf. sericea permixta Williams (p. 133) 

Aldress Shales, Ox Wood Dingle, Shropshire. 
Figs. 2, 10. BB 35552a, b. Latex casts of the exterior and interior of a brachial valve, 
X 6-8. 
Fig. 4. BB 35554b. Latex cast of the exterior of a brachial valve, x 5-1. 
Figs. 6, 5. BB 35553a, b. Latex casts of the exterior and interior of a brachial valve, X 4-8, 

X57- 
Fig. 7. BB 35551a. Internal mould of a pedicle valve, x 3-7. 

Sowerbyella cf. sericea (J. de C. Sowerby) (p. 134) 

Whittery Shales, near Marrington Farm, Shropshire. 
Fig. 11. BB 35555. Internal mould of a pedicle valve, x 3-2. 

Whittery Shales, Whittery Wood, Chirbury, Shropshire. 
Figs. 14, 13. BB 35556a, b. Latex casts of the exterior and interior of a brachial valve, 
X3-5- 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 16. BB 35557. Exterior of the pedicle valve, x 2-7. 

Sericoidea cf. abdita Williams (p. 139) 

Hagley Shales, Church Stoke, Shropshire. 

Fig. 8. BB 35485. Latex cast of the interior of a brachial valve, x 7-7. 

Figs. 15, 12, 9. BB 35484a, b. Latex casts of the exterior and interior and the internal 
mould of a brachial valve, x 8. 

Figs. 18, 17. BB 35487a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x8-6. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 24 




PLATE 25 
Rafinesquina sp. (p. 143) 

Whittery Shales, Whittery Quarry, Shropshire. 
Fig. 1. BB 35496a. Latex cast of the interior of a brachial valve, x 3. 

Murinella sp. (p. 141) 

Meadowtown Beds, Little Weston, Shropshire. 
Figs. 2, 3. BB 35421a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, X5, Xf6. 

Furcitella sp. (p. 140) 

Whittery Shales, Whittery Quarry, Shropshire. 
Figs. 4, 5. BB 35422a, b. Latex cast of the exterior and the internal mould of a brachial 
valve, x 4. 

Rafinesquina delicata sp. nov. (p. 141) 

Meadowtown Beds, Quinton's Quarry, Shropshire. 
Fig. 6. Para type, BB 35562. Latex cast of the interior of a brachial valve, x 2. 
Figs. 8, 7. Paratype, BB 35559a. Latex cast and mould of the interior of a pedicle valve, 
x 2-6. 
Fig. 9. Paratype, BB 35560b. Latex cast of the exterior of a pedicle valve, x 2-9. 
Fig. 10. Paratype, BB 35561a. Internal mould of a pedicle valve, x 2-9. 
Fig. 12. Paratype, BB 35563a. Latex cast of the interior of a brachial valve, x 3-8. 

Meadowtown Beds, near Minicop Farm, Shropshire. 
Fig. 11. Holotype, BB 35558a. Internal mould of a brachial valve, x 2-9. 
Fig. 13. Paratype, BB 35564b. Latex cast of the exterior of a brachial valve, x 5-4. 

Bystromena perplexa gen. et sp. nov. (p. 146) 

Spy Wood Grit, near Middleton Church, Shropshire. 
Fig. 14. BB 35366a. Latex cast of the interior of a pedicle valve, x 5-6. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 25 




PLATE 26 
Rafinesquina sp. (p. 143) 

Whittery Shales, Whittery Quarry, Shropshire. 
Fig. 1. BB 35496b. Latex cast of the exterior of a brachial valve, x 5. 

Kjaerina (Hedstroemina) sp. (p. 145) 

Spy Wood Grit, near Rorrington, Shropshire. 
Fig. 2. BB 35452. Internal mould of a pedicle valve, x i-8. 

Bystromena perplexa gen. et sp. nov. (p. 146) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 4, 3. Holotype, BB 35363. Latex cast and mould of the interior of a pedicle valve, 
X3-4. x 4 . 
Figs. 14, 9. Paratype, BB 35365a. Latex cast and mould of the interior of a pedicle valve, 

X4'5- 

Figs. 10, 5. Paratype, BB 35364a. Latex cast and mould of the interior of a pedicle valve, 
X6-6, X43. 

Figs. 6, 7. Paratype, BB 35370. Latex cast and mould of the interior of a brachial valve, 
X7, X5. 

Spy Wood Grit, Spy Wood Brook, Shropshire. 
Figs, 12, 13. Paratype, BB 35368a. Latex cast and mould of the interior of a brachial 
valve, x 6. 

Spy Wood Grit, near Middleton Church, Shropshire. 
Fig. 8. Paratype, BB 35367a. Latex cast of the interior of a brachial valve, x 6. 
Fig. 11. Paratype, BB 35366a. Internal mould of a pedicle valve, x 7-3. 



Bull. Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 26 


















v- 







PLATE 27 
Bystromena perplexa gen. et sp. nov. (p. 146) 

Spy Wood Grit, near Middleton Church, Shropshire. 
Figs. 2, 1. Paratype, BB 35367a, b. Latex cast of the exterior and the internal mould of 
a brachial valve, x 7-5. 

Rostricellula sparsa Williams (p. 153) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 3, 7. BB 35584a, b. Latex cast of part of the exterior and the internal mould of a 
brachial valve, x 6-9, x 10. 

Fig. 4. BB 35585a. Internal mould of a pedicle valve, x 7-5. 

Leptaena cf. ventricosa Williams (p. 148) 

Whittery Shales, near Hockleton Bridge, Shropshire. 
Fig. 5. BB 35454. Internal mould of pedicle valve, x 2-5. 

Whittery Shales, near Marrington Farm, Shropshire. 
Figs. 8, 9. BB 35453a, b. External mould and latex cast of the interior of a brachial valve, 
X4- x 3-4- 

Kiaeromena cf. kjerulfi (Holtedahl) (p. 149) 

Hagley Volcanics, near Church Stoke Hall, Shropshire. 
Fig. 6. BB 37132a. Latex cast of the interior of a brachial valve, x 4-8. 

Whittery Shales, River Camlad, Shropshire. 
Figs, ii, 10. BB 35406a, b. Latex cast of the ornament and the internal mould of a pedicle 
valve, x 6, X2. 

Zygospira sp. (p. 154) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 13, 12. BB 37133. Latex cast and mould of the interior of a brachial valve, x n. 



Bull. Br. Mm. not. Hist. (Geol.) Suppl. n 

IMS 




&PV JS4 



PLATE 28 
Euorthisina cf. moesta minor Havlicek (p. 150) 

Mytton Flags, near St Luke's Church, Shropshire. 
Fig. 1. BB 35377a. Moulds of a group of shells in an orthocone, x 5. 
Fig. 5. BB 35377b. Latex cast of the exterior of a brachial valve in the orthocone, x 6. 
Fig. 2. BB 35378a. Internal mould of a pedicle valve, X4-4- 

Mytton Flags, near Wood House, Shropshire. 
Fig. 3. BB 35379a. Internal mould of a pedicle valve, X6-3. 
Fig. 8. BB 35380a. Internal mould of a brachial valve, x 4. 

Euorthisina moesta (Barrande) (p. 150) 

Sarka Formation (Llanvirn), Rokycany, Czechoslovakia. 
Fig. 4. BB 37160. Dorsal view of the internal mould of a shell, x i-8. 
Fig. 6. BB37161. Internal mould of a pedicle valve, x 2-4. 
Fig. 7. BB 37162. Internal mould of a brachial valve, x 2-5. 

Parastrophinella musculosa sp. nov. (p. 151) 

Spy Wood Grit, near Rorrington, Shropshire. 
Figs. 9, 17. Holotype, BB 35598a, b. Latex cast of the exterior and the internal mould of a 
pedicle valve, X 5, X 4. 

Fig. 10. Paratype, BB 37108. Internal mould of a brachial valve, x 3-7. 
Fig. 11. Paratype, BB 37107. Internal mould of a pedicle valve, X5"5. 
Fig. 12. Paratype, BB 35599a. Internal mould of a brachial valve, x 5. 
Fig. 13. Paratype, BB 35600. Internal mould of a brachial valve, X4"5- 

Parastrophinella sp. (p. 152) 

Hagley Volcanics, near Church Stoke Hall, Shropshire. 
Figs. 14, 15. BB 35461a, b. Latex cast of the exterior and the internal mould of a pedicle 
valve, x i-8. 

Sericoidea cf. abdita Williams (p. 139) 

Hagley Shales, near Church Stoke, Shropshire. 
Fig. 16. BB 35486b. Internal mould of a brachial valve, xii-5- 



Bull, Br. Mus. nat. Hist. (Geol.) Suppl. n 



PLATE 7.X 




A LIST OF SUPPLEMENTS 

TO THE GEOLOGICAL SERIES 

OF THE BULLETIN OF 

THE BRITISH MUSEUM (NATURAL HISTORY) 



i. Cox, L. R. Jurassic Bivalvia and Gastropoda from Tanganyika and Kenya. 
Pp. 213 ; 30 Plates ; 2 Text-figures. 1965. £6. 

2. El-Naggar, Z. R. Stratigraphy and Planktonic Foraminifera of the Upper 
Cretaceous — Lower Tertiary Succession in the Esna-Idfu Region, Nile Valley, 
Egypt, U.A.R. Pp. 291 ; 23 Plates ; 18 Text-figures. 1966. £10. 

3. Davey, R. J., Downie, C, Sarjeant, W. A. S. & Williams, G. L. Studies on 
Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 248 ; 28 Plates ; 64 Text- 
figures. 1966. £j. 

3. Appendix. Davey, R. J., Downie, C, Sarjeant, W. A. S. & Williams, G. L. 
Appendix to Studies on Mesozoic and Cainozoic Dinoflagellate Cysts. Pp. 24. 
1969. 8op. 

4. Elliott, G. F. Permian to Palaeocene Calcareous Algae (Dasycladaceae) of 
the Middle East. Pp. 111 ; 24 Plates ; 17 Text-figures. 1968. £5.12^. 

5. Rhodes, F. H. T., Austin, R. L. & Druce, E. C. British Avonian (Carboni- 
ferous) Conodont faunas, and their value in local and continental correlation. 
Pp. 315 ; 31 Plates ; 92 Text-figures. 1969. £11. 

6. Childs, A. Upper Jurassic Rhynchonellid Brachiopods from Northwestern 
Europe. Pp. 119 ; 12 Plates ; 40 Text-figures. 1969. £4.75. 

7. Goody, P. C. The relationships of certain Upper Cretaceous Teleosts with 
special reference to the Myctophorids. Pp. 255 ; 102 Text-figures. 1969. 
£6.50. 

8. Owen, H. G. Middle Albian Stratigraphy in the Anglo-Paris Basin. Pp. 164 ; 
3 Plates ; 52 Text-figures. 1971. £6. 

9. Siddiqui, Q. A. Early Tertiary Ostracoda of the family Trachyleberididae 
from West Pakistan. Pp. 98 ; 42 Plates ; 7 Text-figures. 1971. £8. 

10. Forey, P. L. A revision of the Elopiform fishes, fossil and Recent. Pp. 222 ; 
92 Text-figures. 1973. £9.45. 



Printed in Great Britain by John Wright and Sons Ltd. at The Stonebridge Prets, Bristol BS4 5NU